Hyperspectral anomaly detection is a crucial technique for recognizing abnormal pixels in hyperspectral images (HSIs), that is, those with distinct spectral characteristics from those of the surrounding background. Traditional methods always fall short in effectively leveraging the information regarding the spectral and spatial aspects of the dataset simultaneously, limiting their detection performances. This article proposes a novel framework using U-Net, termed hybrid convolution and transformer-based U-Net (HCT-Unet), which integrates convolution with a multihead attention mechanism in Transformer for enhanced hyperspectral anomaly detection. To ensure a more comprehensive understanding of spatial and spectral interactions, the HCT-Unet architecture capitalizes on the strengths of local feature extraction of convolutional layers and the capabilities of the long-range dependency modeling of Transformers. A key innovation of this framework is an error attention mechanism, which facilitates adaptive multiscale feature fusion and enhances the feature representation capacity. Furthermore, a new anomaly score calculation method is proposed, which combines reconstruction error with the pixelwise structural similarity index (SSIM) to determine pixel anomaly from both local structural preservation and global spectral consistency perspectives. Experiments carried out on seven different hyperspectral datasets reveal that the proposed method consistently outperforms the widely accepted state-of-the-art methods in hyperspectral anomaly detection.

Accurate Local Climate Zone (LCZ) classification is essential for urban climate studies, environmental monitoring, and sustainable city planning. Recent advances in Deep Learning (DL) have significantly improved LCZ mapping, but challenges remain in capturing spatial position features and distinguishing spectrally similar land cover types. This letter proposes multi-scale Coordinate Attention-based Multi-Stream Local Climate Zone Network (MSCA-MSLCZNet), a multi-stream DL framework integrating multi-scale feature processing, attention mechanisms, and rule-based refinement to enhance LCZ classification performance. The model is evaluated on the So2Sat LCZ42 dataset, outperforming baseline methods in overall accuracy (OA), OA of built-up classes (OAbu), OA of natural classes (OAn), and Kappa. Further validation on Milan LCZ mapping confirms its generalization capability, demonstrating strong classification performance in built-up areas and improved urban structure delineation. Comparative experiments highlight the model’s ability to better differentiate urban structures and built-up zones from natural landscapes. MSCA-MSLCZNet proves effective for large-scale LCZ mapping, offering improved classification accuracy and adaptability to diverse geographic regions.

Abstract—Coastal water quality monitoring is crucial for environmental sustainability and public health. This work introduces a very cutting-edge methodology, using ΦSat-2 multispectral data and quanvolutional neural networks to explore quantum-enhanced machine learning for water contaminant assessment. By integrating quantum preprocessing into a classical regression model, it is possible to achieve a significant reduction in model parameters while maintaining high predictive accuracy. Additionally, this work introduces an innovative dataset that integrates simulated ΦSat-2 spectral data with Copernicus Marine Service bio-geochemical products, ensuring a strong alignment between satellite observations and reference turbidity measurements. Our results show that quantum models use up to 98% fewer parameters than their classical counterparts, while achieving a 6.9% improvement in the Pearson correlation coefficient between the ΦSat-2 pre-processed bands and the ground-truth turbidity values, compared to the case without quantum pre-processing. Additionally, the Root Mean Square Error (RMSE) improves by 7.3% over the classical baseline. These findings highlight the potential of quantum-assisted remote sensing to enable more efficient and scalable analysis of large-scale water contaminant data, paving the way for advanced big data approaches in water quality monitoring.

Detecting rice transplantation dates is crucial for understanding its effect on grain yield and water consumption at regional scales. Traditionally, identifying the rice transplantation phase using dual-polarized (dual-pol) synthetic aperture radar (SAR) data has relied on backscatter intensity due to its characteristic low values during the flooding stage. This study leverages a recently proposed dual-pol radar surface index (DpRSI) to analyze the spatiotemporal dynamics of the rice transplantation phases. Using this index, we propose an unsupervised framework to identify rice transplantation dates. The framework is evaluated using ground-truth (GT) data over rice-cultivated regions in Vijayawada, India, during the kharif season 2018, demonstrating its effectiveness in detecting shifts in transplantation dates over a large spatial extent.

Hyperspectral unmixing is a technique in hyperspectral image processing that decomposes the spectra of mixed pixels into pure spectral components (endmembers) and their corresponding contributions (abundances). When dealing with complex mixed-terrain scenes, such as urban areas, significant challenges arise due to the complexity of the environment. Urban areas feature intricate geometric structures in individual pixels, including diverse 2-D and 3-D structures and the composite use of various building materials, resulting in highly complex scenarios. To address these challenges, this work exploits urban auxiliary information in the framework of an adaptive multitask autoencoder (AE) unmixing model, utilizing graph associations. The framework enhances the information in hyperspectral images by utilizing urban auxiliary data. Specifically, it performs superpixel segmentation to subdivide complex urban environments into simpler units. Subsequently, different AE-based unmixing methods are applied to these segmented results. Graph associations are employed to identify similar blocks in the image, incorporating this additional information into the unmixing process. In the experiments conducted for this work, two hyperspectral unmixing datasets were prepared, along with their corresponding urban auxiliary data. The results demonstrate that the proposed method achieves robust performance, even in complex urban environments.

Presents corrections to the paper, (Corrections to “Adaptive Multitask Autoencoder-Based Hyperspectral Unmixing Exploiting Auxiliary Data via Graph Associations”).

A significant amount of remotely sensed data is generated daily by many Earth observation (EO) spaceborne and airborne sensors over different countries of our planet. Different applications use those data, such as natural hazard monitoring, global climate change, urban planning, and more. Many challenges are brought by the use of these big data in the context of remote sensing (RS) applications. In recent years, the employment of machine learning (ML) and deep learning (DL)-based algorithms has allowed a more efficient use of these data, but the issues in managing, processing, and efficiently exploiting them have even increased as classical computers have reached their limits. This article highlights a significant shift toward leveraging quantum computing (QC) techniques in processing large volumes of RS data. The proposed Quanv4EO framework introduces a quanvolution method for (pre)processing multidimensional EO data. Its effectiveness was first demonstrated on standard image classification datasets (MNIST and FashionMNIST), achieving accuracies of 99.84% and 96.81%, respectively, with a significantly reduced model size of 42 k parameters and 16 frozen qubits. Its capabilities were then checked on EO datasets, such as EuroSAT, with a mean accuracy of 96% using balanced iterative reducing and clustering using hierarchies (BIRCHs) clustering and 93% using automated DL (AutoDL), surpassing or matching state-of-the-art (SOTA) classical nonquantum models. Applying the framework to synthetic aperture radar (SAR) data, the QSPeckleFilter demonstrates notable improvements in speckle noise reduction, achieving a peak signal-to-noise ratio (PSNR) of 21.72 and a structural similarity index measure (SSIM) of 0.81, surpassing all tested classical counterparts. The proposed results underscore the potential of quantum-enhanced approaches in RS data analysis, paving the way for more efficient and effective solutions for wide geographical area EO data exploitation.

Synthetic aperture radar (SAR) data classification has gained significant research interest, as accurate land-cover information is vital in a wide range of planning and management activities. While classification algorithms for full-polarimetric (full-pol) SAR data are typically based on the statistical or physical characteristics of the scattering mechanism from targets, classification of co-cross polarization (VV-VH or HH-HV) dual-polarimetric (dual-pol) SAR data has traditionally relied on backscatter intensity information due to its limited polarimetric information. Several studies also employ the dual-pol entropy/alpha decomposition parameters, establishing a conventional framework for supervised and unsupervised classification of dual-pol SAR data. However, it is essential to note that the conventional approach cannot differentiate between certain elementary targets, leading to misclassification among diverse land-cover targets. To address this limitation, we introduce an unsupervised clustering technique for dual-pol Sentinel-1 SAR data utilizing the conventional entropy parameter alongside a dual-pol target characteristic parameter that discriminates between various land-cover targets, including “dihedral-like” (buildings, etc.) and “surface-like” (water bodies, etc.) targets in a dual-pol scene. Thus, the proposed clustering scheme, which applies to both single look complex (SLC) and ground range detected (GRD) SAR, categorizes it into eight clusters, each representing specific target characteristics. We adopted two strategies to assess the proposed clustering scheme: 1) cluster zones obtained for diverse land-cover targets spanning continents and 2) temporal changes in cluster zones over rice-cultivated fields at various growth stages. The proposed approach effectively discriminates diverse land-cover targets and distinct growth stages of rice.

Urban hyperspectral image super-resolution can rapidly acquire high-quality data with rich spatial details and spectral fidelity through technical means for urban development. However, conventional super-resolution methods for multi-spectral and natural images struggle to meet the aforementioned data requirements simultaneously. this paper proposes an urban hyperspectral super-resolution method that combines subpixel mapping and interpolation. This method aims to ensure spectral information and minimize time consumption through a high-precision surface interpolation method based on curve theory. Additionally, by utilizing subpixel mapping to introduce urban unmixing information. Finally, by employing wavelet transformation, the method integrates the effective information from both approaches, obtaining urban hyperspectral images with enhanced spatial detail and spectral fidelity. This method has been subjected to comprehensive experimentation, affirming that our proposed method surpasses the current state-of-the-art super-resolution in terms of performance and effectiveness.

Gated Neighborhoods are residential areas characterized primarily by being closed urbanized spaces, with low construction density, and low-rise buildings, as well as significant areas dedicated to green spaces and sometimes to bodies of water. They are widely spread in large urban areas across South America, and they can be found in many locations within the Metropolitan Area of Buenos Aires, Argentina.Despite their significant expansion in recent years, public geographic database of GNs in Argentina is outdated, making it difficult to monitor their expansion over natural areas such as wetlands. To automate the process of mapping GNs we developed a methodology using PlanetScope images in input and the YOLOv5 convolutional neural network as a processing tool. The results show Precision of 0.85, Recall of 0.71 and mean IoU of 0.8. These are promising results to generate an operative tool that could aid in the procedure of mapping GNs.

The use of Synthetic Aperture Radar (SAR) has greatly advanced our capacity for comprehensive Earth monitoring, providing detailed insights into terrestrial surface use and cover regardless of weather conditions, and at any time of day or night. However, SAR imagery quality is often compromised by speckle, a granular disturbance that poses challenges in producing accurate results without suitable data processing. In this context, the present paper explores the cutting-edge application of Quantum Machine Learning (QML) in speckle filtering, harnessing quantum algorithms to address computational complexities. We introduce here QSpeckleFilter, a novel QML model for SAR speckle filtering. The proposed method compared to a previous work from the same authors showcases its superior performance in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM) on a testing dataset, and it opens new avenues for Earth Observation (EO) applications.

This paper explores an innovative fusion of Quantum Computing (QC) and Artificial Intelligence (AI) through the development of a Hybrid Quantum Graph Convolutional Neural Network (HQGCNN), combining a Graph Convolutional Neural Network (GCNN) with a Quantum Multilayer Perceptron (MLP). The study highlights the potentialities of GCNNs in handling global-scale dependencies and proposes the HQGCNN for predicting complex phenomena such as the Oceanic Niño Index (ONI). Preliminary results suggest the model potential to surpass state-of-the-art (SOTA). The code will be made available with the paper publication.

Climate change is intensifying extreme weather events, causing both water scarcity and severe rainfall unpredictability, and posing threats to sustainable development, biodiversity, and access to water and sanitation. This paper aims to provide valuable insights for comprehensive water resource monitoring under diverse meteorological conditions. An extension of the SEN2DWATER dataset is proposed to enhance its capabilities for water basin segmentation. Through the integration of temporally and spatially aligned radar information from Sentinel-1 data with the existing multispectral Sentinel-2 data, a novel multisource and multitemporal dataset is generated. Benchmarking the enhanced dataset involves the application of indices such as the Soil Water Index (SWI) and Normalized Difference Water Index (NDWI), along with an unsupervised Machine Learning (ML) classifier (k-means clustering). Promising results are obtained and potential future developments and applications arising from this research are also explored.

Synthetic Aperture Radar (SAR) data, particularly for Asian countries, are valuable in monitoring crops. Scattering information extracted from full-polarimetric (FP) SAR data offers exceptional sensitivity to crop water content and geometrical properties. Thus, they are widely used for continuous crop monitoring throughout their growth stages. However, many of these methods are limited to FP SAR data. This study introduces a new approach to obtain scattering power components from SLC and GRD dual-polarimetric (DP) SAR data. We found the scattering powers obtained from the dual-pol SAR data to be sensitive to changes in the crop morphology as it progresses to advanced growth stages.

Prospective urban growth models serve as tools aimed at supporting sustainable city development, informed decision-making and proactive planning, contributing to enhancing livability and resilience in expanding urban areas.This research develops a prospective urban growth model for the Metropolitan Area of Córdoba (Argentina) by integrating multispectral remote sensing data and open-source GIS data within a Deep Learning framework, projecting the urban extent for the year 2030. It is observed that the DL model effectively replicates edge expansion by utilizing neighborhood information. Additionally, it accurately simulates outlying growth via isolated patches, all of which follow similar patterns to those observed in the past.

LoRaWAN technology’s reliability is challenged by weather parameters, which can influence the communication channel design, especially when dealing with outdoor devices. We propose to analyze this effect by evaluating the relationship between the received signal strength indicator (RSSI) and different weather parameters, as well as its temporal changes. A rigorous statistical analysis of the RSSI sequences is conducted to assess if they could be represented by a specific statistical model. For this purpose, several models are investigated. The Artificial Intelligence (AI) algorithms cover machine learning (ML) and deep learning methods are appealing when dealing with time series forecasting. Nevertheless, the classical autoregressive integrated moving average (ARIMA) model can be an attractive alternative due to its simplicity. Therefore, this work proposes a comparative study of ARIMA, AI, and hybrid approaches to forecast the RSSI using weather parameters as regressors. The considered AI algorithms are the artificial neural network (ANN), support vector machine (SVM), random forest (RF), and Long Short-Term Memory (LSTM). Also, hybrid models are constructed, coupling the ARIMA with them. The models are evaluated in time series of RSSI, measured by eight different LoRaWAN transmitter nodes and considering the temperature, pressure, relative humidity, and rain as weather parameters. Our analysis reveals that temperature is the dominant factor among weather parameters, and negatively affects RSSI. The ARIMA model that uses only the temperature as a regressor provides consistently better fits than the ARIMA without regressors. Moreover, coupling the ARIMA with the temperature as a regressor and the ANN (ARIMA-ANN) is the best option among the pure AI and hybrid approaches. However, it provided accuracy measures very close to those obtained from the ARIMA model fitted in the first stage, with similar performance. Therefore, the ARIMA model considering the temperature is the most competitive alternative when analyzing RSSI measurements, with the advantage of being the most straightforward method. These results suggest that the RSSI from the analyzed LoRaWAN receiver nodes may not present nonlinear patterns and, considering several weather parameters, they are affected mainly by the outdoor temperature.

Spaceborne temporal sequences of synthetic aperture radar (SAR) data have a definite advantage over multispectral data sequences in terms of continuity and regularity. Still, deep-learning (DL) applications in remote sensing have primarily focused on multispectral data. This work is focused instead on a novel 3-D DL architecture for SAR data sequences. The proposed approach utilizes a trained-from-scratch 3-D fully convolutional network (FCN) with a 3-D ResNet-50 as a backbone to classify ten land cover types using multitemporal Sentinel-1 SAR data. Experimental results show that this architecture provides a trained model that outperforms existing DL methods applied to the same SAR sequence in terms of overall accuracy (OA). In addition, the results using only SAR data provide very similar and consistent performances to those achievable using multispectral data. Accordingly, the proposed approach demonstrates the potential of SAR temporal sequences in land cover mapping using DL techniques.

The foundation model has recently garnered significant attention due to its potential to revolutionize the field of visual representation learning in a self-supervised manner. While most foundation models are tailored to effectively process RGB images for various visual tasks, there is a noticeable gap in research focused on spectral data, which offers valuable information for scene understanding, especially in remote sensing (RS) applications. To fill this gap, we created for the first time a universal RS foundation model, named SpectralGPT, which is purpose-built to handle spectral RS images using a novel 3D generative pretrained transformer (GPT). Compared to existing foundation models, SpectralGPT 1) accommodates input images with varying sizes, resolutions, time series, and regions in a progressive training fashion, enabling full utilization of extensive RS Big Data; 2) leverages 3D token generation for spatial-spectral coupling; 3) captures spectrally sequential patterns via multi-target reconstruction; and 4) trains on one million spectral RS images, yielding models with over 600 million parameters. Our evaluation highlights significant performance improvements with pretrained SpectralGPT models, signifying substantial potential in advancing spectral RS Big Data applications within the field of geoscience across four downstream tasks: single/multi-label scene classification, semantic segmentation, and change detection.

Hyperspectral target detection aims to locate targets of interest in the scene, and deep learning-based detection methods have achieved the best results. However, black box network architectures are usually designed to directly learn the mapping between the original image and the discriminative features in a single data-driven manner, a choice that lacks sufficient interpretability. On the contrary, this article proposes a novel deep spatial-spectral joint-sparse prior encoding network (JSPEN), which reasonably embeds the domain knowledge of hyperspectral target detection into the neural network, and has explicit interpretability. In JSPEN, the sparse encoded prior information with spatial-spectral constraints is learned end-to-end from hyperspectral images (HSIs). Specifically, an adaptive joint spatial-spectral sparse model (AS2JSM) is developed to mine the spatial-spectral correlation of HSIs and improves the accuracy of data representation. An optimization algorithm is designed for iteratively solving AS2JSM, and JSPEN is proposed to simulate the iterative optimization process in the algorithm. Each basic module of JSPEN one-to-one corresponds to the operation in the optimization algorithm so that each intermediate result in the network has a clear explanation, which is convenient for intuitive analysis of the operation of the network. With end-to-end training, JSPEN can automatically capture the general sparse properties of HSIs and faithfully characterize the features of background and target. Experimental results verify the effectiveness and accuracy of the proposed method. Code is available at https://github.com/Jiahuiqu/JSPEN.

Hyperspectral sensors can rapidly acquire high-quality spectral data, very useful for urban monitoring applications. Unfortunately, their spatial detail is not fine enough, and methods to enhance this resolution are required. However, conventional super-resolution (SR) methods for multispectral data do not match the requirements needed to maintain high spectral fidelity. Therefore, this article proposes a hyperspectral SR method that combines subpixel mapping and interpolation, and whose main aim is to enhance urban monitoring. This method aims to guarantee spectral quality and minimize computational time through a high-precision surface interpolation method based on curve theory. Moreover, unmixing-based subpixel mapping is exploited to introduce unmixing information. Finally, using wavelet transforms, the method integrates the effective information from the two previous approaches, obtaining urban hyperspectral images with enhanced spatial details and spectral fidelity. This method has been subjected to a comprehensive experimentation, affirming that the proposed method surpasses the current state-of-the-art SR results in terms of performance and effectiveness.

In the remote sensing field, detecting small objects is a pivotal task, yet achieving high performance in deep learning-based detectors heavily relies on extensive data annotation. The challenge intensifies as small objects in remote sensing imagery are typically densely distributed and numerous, leading to a substantial increase in the cost of creating large-scale annotated datasets. This elevated cost poses significant limitations on the application and advancement of small object detection. To address this issue, a point-based annotation (PBA) method is proposed, which generates bounding boxes (BBOXs) through graph-based segmentation. In this framework, user annotations categorize nodes into three distinct classes—positive, negative, and to-cut—facilitating a more intuitive and efficient annotation process. Utilizing the max-flow algorithm, our method seamlessly generates oriented BBOXs (OBBOXs) from these classified nodes. The efficacy of PBA is underscored by our empirical findings. Notably, annotation efficiency is enhanced by at least 40%, a significant leap forward. Moreover, the intersection over union (IoU) metric of our OBBOX outperforms existing methods like “segment anything model (SAM)” by 10%. Finally, when applied in training, models annotated with PBA exhibit a 3% increase in the mean average precision (mAP) compared with those using traditional annotation methods. These results not only affirm the technical superiority of PBA but also its practical impact on advancing small object detection in remote sensing.

Change detection is a fundamental task that involves assessing changes in a given region over multiple time periods. It has been widely applied across various fields, including monitoring deforestation, urban expansion, and natural disaster analysis. In this article, we address the critical and complex issue of automatically identifying types of changes in land cover using remotely sensed imagery. While conventional unsupervised change detection methods typically focus on comparing pairs of images and making a binary decision between “change” and “nonchange,” our approach tackles the challenge of analyzing long image series and identifying the kind of change. Under this condition, the unsupervised change detection process allows for a more informative identification of the land cover dynamics. Moreover, our approach transforms input data to a new representation, capturing the target’s spectral response changes over time. Through the utilization of stochastic distances and an optimized thresholding scheme, areas exhibiting minimal spectral response variance are classified as unchanged, effectively distinguishing them from regions undergoing modifications. Next, by applying autocorrelation analysis, regions exhibiting temporal modifications are segregated into periodic (i.e., seasonal) and aperiodic (i.e., permanent) change cases. Experimental validation using both simulated and real-world remote sensing image series demonstrates the effectiveness of the proposed approach.

Target characterization parameters are pivotal in accurately identifying and assessing diverse land cover targets in radar polarimetry. While full-polarimetric (full-pol) synthetic aperture radar (SAR) data offer numerous parameters, characterizing targets with HH-HV or VV-VH dual-polarimetric (dual-pol) SAR data has traditionally relied on backscatter intensity alone due to limited polarimetric information, which leads to ambiguities in characterizing diverse land cover targets. In response to this limitation, this study introduces a novel target characteristic parameter$\overline {\alpha }_{(k)}$derived from dual-pol single-look complex (SLC) and ground range detected (GRD) SAR data that are capable of discriminating between “dihedral-like” (buildings, bridges, ships, and so on) and “surface-like” (water bodies, bare fields, runways, and so on) targets, by employing a data-driven approach. We first derive a set of normalized descriptors independently of SLC and GRD SAR data to formulate two indices that characterize “dihedral-like” and “surface-like” targets. Using the two indices, we derive the dual-pol target characteristic parameter, providing a novel perspective on the intricate nature of radar responses from diverse land cover targets acquired by dual-pol SAR sensors. Furthermore, we employ this parameter to extract three scattering power components: “dihedral-like” ($P_{d-l}$), unpolarized ($P_{u}$), and “surface-like” ($P_{s-l}$) from both dual-pol SLC and GRD SAR data. We assess the proposed target characteristic parameter and scattering power components using Sentinel-1 images acquired over diverse land cover targets spanning six continents. This novel approach enables improved global land cover characterization with operational SAR missions such as Sentinel-1 and upcoming NASA-ISRO SAR (NISAR) missions.

Multitask learning has been widely applied in visual learning to significantly enhance the performance. The combination of hyperspectral change detection (HCD) and band reweighting can achieve discriminative feature enhancement for improving detection performance. However, existing multitask models for these two tasks are unidirectional, with band reweighting unable to learn from task guidance. To address this challenge, a multitask HCD (MHCD) framework with differential band reweighting and unbalanced contrastive learning is proposed. MHCD consists of a differential band reweighting network (DBRN) and a Siamese detection network. DBRN extracts discriminative information for HCD by analyzing the differential spatial-spectral information across time states, whose optimization is under the guidance of HCD. Furthermore, a multitemporal interaction module and multidomain fusion module are inserted into the Siamese detection network. They hierarchically connect cross-temporal features and fuse features from spatial, spectral, and temporal domains, providing complementary clues in these different domains. Considering the sample imbalance and enormous variation within a class in binary HCD, an unbalanced contrastive learning method based on multiple prototypes (UCLM) tailored has been considered. It estimates multiple prototypes to flexibly adjust the contribution of different classes of samples to the loss. The proposed method has been validated using three public benchmark datasets, demonstrating improvements in multiple metrics for change detection. The code of our paper is available at:https://github.com/jiefeng0109/MHCD.

The previously presented self-organizing pixel entanglement neural network (SOPENN) model only establishes 2-D basis vectors that are orthogonal to each other in the Hilbert space, which cannot sufficiently reflect the spectral information and the entanglement characteristics of pixels in multispectral images. Therefore, an iterative self-organizing pixel matrix entanglement (ISOPME) image classification model is proposed in this article. Quantum pixel matrix entanglement (PME) is based on quantum pixel entanglement, which considers a pixel as a quantum, and the quantum entanglement theory in quantum informatics is applied in the PME. First, the PME theory was developed to associate the quantum states of pixels with their gray values. Second, the PME coefficient was proposed to extract the entanglement relationship between pixel matrices with 3-D basis vectors in the Hilbert space. Finally, an ISOPME model was developed to implement a self-organizing clustering for multispectral remote sensing image classification. The experimental results for four test areas demonstrate that: 1) the proposed ISOPME approach achieves an average classification accuracy of 92.02% and a Kappa coefficient (KC) of 0.88; 2) when compared with four traditional unsupervised classification methods, ISOPME on average improves the classification accuracy by 8.96% and the KC by 0.14; and 3) the classification accuracy and KC from ISOPME reach the same level as the more sophisticated supervised classification methods, such as support vector machine (SVM), and are close to those obtained using deep learning (DL) classification methods.

Unlike periodic changes in natural cover, urban construction activities caused by urbanization show a distinctly non-periodic pattern in time. It is desirable to capture and recognize these changes, to timely update urban information databases among the others, by utilizing high-frequency observation of optical sensors or synthetic aperture radar (SAR) in an automatic way. To this aim, the primary task is to segment long time sequences and distinguish between changing and non-changing time segments. Unfortunately, urban building activities have different durations. Following up our previous work of monitoring urban building construction activities by using SAR coherent time series data [6], in this paper we focus on distinguishing among changed segments of different duration by a proposed semantic segmentation method based on LSTM autoencoders.

Monitoring the United Nation Sustainable Development Goals (UN-SDGs) calls for adequate methodologies to extract specific indicators of status for each realm (e.g., air, water, land) and socio-economic-environmental issue. Earth observation has been considered, since the beginning, one of the pillars of this task, but the use of Synthetic Aperture Radar is still limited. This work report about the results achieved by a joint research project between Italy and Argentina, whose goal was to design and implement data processing techniques able to exploit the data from the COSMO-SkyMed and SAOCOM constellations, in conjunction with those by the Sentinel constellation by the European Space Agency and provide country-wide indicators for some of the UN-SDGs.

Multitemporal satellite images can be represented as a three-dimensional cube. This remote sensing data type require modeling techniques comprising spatial and temporal dependence altogether. This work aims at developing a hybrid framework combining the three-dimensional autoregressive (3D-AR) statistical model and machine learning algorithms to accommodate spatial and temporal correlations in a stack of temporal synthetic aperture radar (SAR) images. We propose using the 3D-AR to model the linear dependence among the voxels and one among artificial neural networks (ANN), support vector regression (SVR), and random forest (RF) for modeling the nonlinear component. The resulting models, labeled as 3D-AR-ANN, 3D-AR-SVR, and 3D-AR-RF hybrid models, respectively, can help to predict missing voxels or to forecast a one-step-ahead image of a remote sensing data sequence in the presence of linear and nonlinear patterns. The procedure is validate by applying the proposed hybridization approach to predict different parts of a SAR image stack from the urban area of Cordoba, Argentina. The results show the usefulness of our approach in predicting multitemporal SAR data, being more accurate than the state-of-the-art 3D-AR model.

Precipitation is critical to human life, especially in developing nations like Ethiopia, where rain-fed agriculture supports the economy and food security. High variability and limited precipitation data are critical problems in the Awash River Basin. In this regards, high-resolution satellite-based precipitation estimation can fulfill these data gaps. In this study, ERA5, CHIRPS, IMERG-V06, and PERSIANN-CDR were compared with gauged observations from 24 meteorological stations between 2001 and 2013.Using pixel-to-point pair-wise comparison, systematic shortcomings in satellite rainfall estimates were identified. Accordingly, CHIRPS outperform with R2values from 0.95 at Bole to 0.48 at Assaita and RMSE values from 57.82 mm at Tulu Bolo to 28.82 mm at Dupity whereas PERSIANN-CDR and ERA5 poorly performed. The performance of each products declines from highlands to lowlands. This suggests that topographic effects should be considered when using these products for hydrologic, climate change, and extreme event forecasting.

Food traceability in organic agriculture requires a comprehensive "crop history" that includes information from the moment seedlings begin to sprout. Radar remote sensing could contribute in this framework by providing satellite-observable variables and time sequences that help build a more complete crop history. One possible application of this concept is monitoring tillage techniques, which have different impacts on soil properties. The so-called "minimum tillage" reduces erosion and surface runoff, and translates into different backscattering mechanisms in spaceborne radar observation, compared to those of conventional ploughing. In this work, preliminary experiments were conducted to assess the type of tillage based on sequences of spaceborne radar data, specifically plowing-and-harrowing versus minimum tillage. By using radar remote sensing, customers can achieve a more comprehensive understanding of the history of their food from "farm-to-fork", which can be particularly important for high-tier organic food.

Quantum Machine Learning (QML) is an emerging technology that only recently has begun to take root in the research fields of Earth Observation (EO) and Remote Sensing (RS), and whose state of the art is roughly divided into one group oriented to fully quantum solutions, and in another oriented to hybrid solutions. Very few works applied QML to EO tasks, and none of them explored a methodology able to give guidelines on the hyperparameter tuning of the quantum part for Land Cover Classification (LCC). As a first step in the direction of quantum advantage for RS data classification, this letter opens new research lines, allowing us to demonstrate that there are more convenient solutions to simply increasing the number of qubits in the quantum part. To pave the first steps for researchers interested in the above, the structure of a new hybrid quantum neural network for EO data and LCC is proposed with a strategy to choose the number of qubits to find the most efficient combination in terms of both system complexity and results accuracy. We sampled and tried a number of configurations, and using the suggested method we came up with the most efficient solution (in terms of the selected metrics). Better performance is achieved with less model complexity when tested and compared with state-of-the-art (SOTA) and standard techniques for identifying volcanic eruptions chosen as a case study. Additionally, the method makes the model more resilient to dataset imbalance, a significant problem when training classical models. Lastly, the code is freely available so that interested researchers can reproduce and extend the results.

Convolutional neural networks (CNNs) have a strong capacity to extract deep-level features from data. However, the standard convolution (SC) only considers the intensity-information and ignores the spatial gradient-information. Since spatial difference features are more robust to illumination invariance, this letter proposes a Multi-Scale Central Differential Convolutional (MSCDC) network for hyperspectral anomaly detection. Specifically, we use Central Difference Convolution (CDC) to combine intensity- and gradient-information. This solution improves the representation ability of HSIs and enhances the difference between the background and the anomalies. Furthermore, to fully utilize local spatial information and adapt to targets with different sizes, CDC kernels of three different sizes are used to capture high-, mid- and low-level features, respectively. Finally, a SC is used to fuse multi-scale features and obtain more reliable spatial information. Compared with five popular hyperspectral anomaly detection methods on four real-world HSI datasets, the proposed MSCDC exhibits excellent performances.

Hyperspectral unmixing is a crucial task in hyperspectral image processing and analysis. It aims to decompose mixed pixels into pure spectral signatures and their associated abundances. However, most current unmixing methods ignore the reality that the same pixel of a hyperspectral image has many different reflections simultaneously. To address this issue, we propose a multi-task autoencoding model for multiple reflections, which can improve the algorithm’s robustness in complex environments. Our proposed framework uses 3D-CNN-based networks to jointly learn spectral-spatial priors and adapt to different pixels by complementing the advantages of other unmixing methods. The proposed method can quantitatively evaluate each area of data, which helps improve the algorithm’s interpretability. This paper presents MAHUM (Multi-tasks Autoencoder Hyperspectral Unmixing Model), which stacks multiple models to deal with various reflections of complex terrain. We also perform sensitivity analysis on some parameters and show experimental results demonstrating our method’s ability to express the adaptability of different materials in different methods quantitatively.

Spectral distortion severely limits detection performance in hyperspectral imagery, while feature learning with neural networks could provide sufficient capacity to enhance spectral consistency. This paper designs an end-to-end hyperspectral target detection (HTD) network based on transfer learning and nonlinear spectral synthesis (TLNSS). We first utilize bilinear mixture model (BMM) to synthesize nonlinear target and background spectra for training sample augmentation, which could better characterize ground objects in complex environments. Due to the mutual constraints between the quantity and diversity of the synthesized spectra, transfer learning is introduced to further address data insufficiency. Specifically, we propose an asymmetric autoencoder with a particularly designed multi-level loss to maximally distinguish the reconstruction residuals of background and target, where the multi-scale feature extraction sub-network is trained with abundant reference data, and the simple restoration sub-network is updated with the simulated spectra. To effectively reconstruct the input as expected, the features extracted from different blocks are complementarily integrated through residual attention. Lastly, we accumulate reconstruction residuals across all levels for final detection. The experimental results and ablation analysis of single-data detection on three hyperspectral images verify the superiority and effectiveness of the proposed method, and further cross-data detection consolidates the satisfactory tolerance of TLNSS to spectral variation.

Autoencoder (AE) has been widely used in the field of hyperspectral anomaly detection. It is assumed that the background can be reconstructed well, but the anomalies cannot. Hence, the pixels with larger reconstruction error are considered as anomalies. However, owing to the strong nonlinear representation ability of AE, it is difficult to distinguish between background and anomalies. To address this problem, we propose a Residual Self-Attention-based AutoEncoder (RSAAE) for hyperspectral anomaly detection. RSAAE consists of dense residual self-attention modules, an encoder, and a decoder. First, a novel residual self-attention module is designed, which can effectively extract the main features and weaken the ability of subsequent network to reconstruct anomalies, as well as preserve the original features to avoid the deterioration of network performance after the use of dense self-attention modules. Furthermore, inspired by manifold learning, we assume that the background is low-rank in the original space, and has the same property in the latent space after dimensionality reduction. We proposed a low-rank loss function to constrain the latent space, thereby suppressing anomaly reconstruction. Experiments on four real hyperspectral image (HSI) datasets showed that the proposed RSAAE method can produce more accurate detection results than eight popular methods.

The use of multitemporal SAR sequences is becoming more intensive in mapping and change detection applications, be-cause of the availability of data sets with enough time length and sampling frequency. However, this procedure often requires to consider sequences of heterogeneous SAR data sets acquired by different sensors with different spatial resolution, frequency and polarimetric features. Their full exploitation is still an open and interesting research task. This work introduces first a pre-processing sequence ap-plied by the team of the project MultiBiGSARData to obtain calibrated and co-registered heterogeneous SAR sequences. Then, a first example of their use for land cover classification is provided. The results are shown for test areas in Italy and Argentina, because the data sets are obtained by combining images acquired by the SIAGSE constellation, composed by the COSMO-SkyMed (CSK), COSMO-SkyMed Second Generation (CSG), and SAOCOM satellites.

Due to the side effects of asbestos on human health and environments, Italy has banned the use of asbestos-containing materials since 1992, but there are still illegal products with asbestos in daily life. In order to investigate the distributions of asbestos to facilitate its removal, this paper carries out asbestos identification with hyperspectral (HS) and panchromatic (PAN) data captured by the PRISMA satellite over Pavia, Italy. In this work, a pansharpening method with guided filter was used to inject more spatial details from 5m PAN to 30m HS. Then, the possible location of asbestos could be obtained by a bilinear sparse target detector (BSTD). Detection maps using BSTD are compared with that obtained by hierarchical constrained energy minimization (hCEM), ensuring the accuracy and reliability, also compared with the results using matched subspace detector with interaction effects (MSDinter) and adaptive MSD (AMSD) to verify the superiority of the bilinear sparse model.

Built-up area (BA) mapping is vital for understanding the effect of the urban regions on the environment, thereby supporting sustainable development. This study proposes a new dual-pol radar built-up area index (DpRBI) to detect the BA using Sentinel-1 SAR data. The DpRBI formulation is based on the three Stokes vector elements of the scattered wave derived from the 2 × 2 covariance matrix C2. This study uses the Sentinel-1 SAR data sets over Milan, Italy and Barcelona, Spain to map the BA using DpRBI. The overall accuracy of the BA extracted using the proposed technique was found to be 84.19% and 87.95% over Milan and Barcelona, respectively. It is noteworthy that even relatively small low-density BA is precisely classified using the proposed built-up index.

Due to the side effects of asbestos on human health and environments, many countries have banned the use of asbestos-containing materials, but there are still illegal products with asbestos in daily life. In order to investigate the distributions of asbestos to facilitate its removal, this paper studies the feasibility of asbestos identification with HyperSpectral (HS) and panchromatic (PAN) data, taking images captured by the PRISMA and ZY1E 2D satellites over Pavia, Italy as examples. In this work, a pansharpening method with guided filter was used to improve HS image quality in terms of spectral fidelity and spatial details. Then, the possible location of asbestos could be obtained by a nonlinear target detector named BSTD. Considering high computational cost for large-scale remote sensing data processing, we further develop BSTD to its parallelized version (denoted as PBSTD). Given the groundtruth of asbestos over Pavia by the Regional Environmental Protection Agency-ARPA Lombardia, our PBSTD and several popular methods are evaluated from both qualitative and quantitative perspectives, showing that most algorithms could correctly detect large-size asbestos roofs, and the nonlinear PBSTD and MSDinter perform better in small-size asbestos identification than other linear detectors. However, the detection accuracy on small-size asbestos is insufficient in practical applications, which indicates that there are still issues to achieve accurate small-size asbestos identification using coarse-spatial-resolution spaceborne remote sensing.

The binary hypothesis testing (BHT) is one of the most important models in hyperspectral target detection (HTD). However, this model is generally based on a linear mixture model (LMM) and might be inaccurate to reflect target and background characterizations in some scenes. This article presents a bilinear sparse target detector (BSTD) by applying the bilinear sparse mixture model (BSMM) to a popular BHT-based detection algorithm termed adaptive matched subspace detector (AMSD), which takes bilinear target–background interaction and sparse abundance into account. Moreover, as AMSD relies heavily on background subspace, we design a robust background subspace construction method. Specifically, we first classify each pixel into noise, border, or other particular instances according to its density, which is measured by jointly spatial–spectral distance. With the coarse classification map, a class-guided automatic background generation (CABG) process is introduced to reliably generate pure background samples. Detection statistics and component analysis on five real-world hyperspectral images verify the effectiveness of our BSTD method.

Hyperspectral imagery with very high spectral resolution provides a new insight for subtle nuances identification of similar substances. However, hyperspectral target detection faces significant challenges of intraclass dissimilarity and interclass similarity due to the unavoidable interference caused by atmosphere, illumination, and sensor noise. In order to effectively alleviate these spectral inconsistencies, this paper proposes a novel target detection method without strict assumptions on data distribution based on an unconstrained linear mixture model and deep learning. Our proposed detector firstly reduces interference via a specifically designed deep-learning-based hierarchical denoising autoencoder, and then carries out accurate detection with a two-step subspace projection, aiming at background suppression and target enhancement. Additionally, to generate representative background and reliable target samples required in the detection procedure, an efficient spatial-spectral unified endmember extraction method has been developed. Performance comparison with several state-of-the-art detection methods and further analysis on four real-world hyperspectral images demonstrate the effectiveness and efficiency of our proposed target detector.

Built-up area extraction from remote sensing images is essential for urban planning, disaster management and industrial development. In this study, we propose two built-up area indices for full (FP) and dual (DP) polarimetric Synthetic Aperture Radar (SAR) data. The built-up area index for FP SAR data is based on the dominant scattering mechanism of the electromagnetic (EM) waves from urban targets. In contrast, the built-up area index for DP SAR data is based on the scattering reflection symmetry property. The two proposed indexes are validated with full and extracted dual pol (VV-VH) scenes of a C-band RADARSAT-2 SAR data over urban San-Francisco. They show encouraging results in detecting urban areas within a SAR resolution cell. The overall accuracy of delineating built-up area is 84.2% for FP SAR data and 79% for DP SAR data.

The vast amount of spectral information provided by hyperspectral images can be useful for different applications. However, the presence of redundant bands will negatively affect application performance. Therefore, it is crucial to select a relevant subset that preserves the information of the original set. In this paper, we present an automatic and accurate band selection method based on Graph Laplacians. Unlike existing band selection methods, this method exploits two similarity measures simultaneously. Furthermore, it is performed on a superpixel level, so it allows us to preserve not only global but contemporaneously local particularities of original data. Experiments show the importance of measuring the relevance of the bands at local and global scales and the ability of the method to minimize intercorrelation among selected bands, hence improving the selection of the most informative spectral channels.

Within the European Space Agency (ESA) Climate Change Initiative (CCI) project framework it is fundamental to generate High Resolution (HR) annual surface water maps. In this work we present an innovative approach aimed at this task using multi-temporal Sentinel-1 SAR data. Mapping water bodies with dual-polarized radar data everywhere in the world is challenging, as the dual-pol backscatter intensity signal is strongly affected by many factors such as terrain an acquisition geometry. In this study, an existing Medium Resolution Land Cover (MRLC) map is ecploited to automatically collect sample points and build a k-means unsupervised model. A qualitative analysis is first performed in three test areas. Preliminary quantitative analysis is then presented, showing 97% extraction accuracy.

Here, we present a method for temporal modeling of the oviposition activity of Ae. aegypti mosquitoes based on a weighted generalised linear model (GLM) with explanatory environmental effects extracted from freely available remotely sensing (satellite) images. Our results show potential for operational applications. Experimental results are provided using field collected Ae. aegypti eggs count data in Córdoba, Argentina.

Satellite images from the same scene observed over time can be composed in an image stack, which could be modeled as a 3-D cube. To handle this type of remote sensing data, on the one side, unidimensional dynamical models have been considered, modeling each pixel separately along the time (pixel-based approach), and exploring the temporal correlation. On the other side, 2-D approaches have been considered to process each image at one date, exploring the spatial correlation. In this article, we propose a new 3-D autoregressive (AR) (3-D-AR) model useful for multitemporal image interpretation exploring the correlation in three dimensions altogether. The 3-D-AR model is statistically defined, and a robust parameter estimation method is discussed. The tools for filtering, forecasting, and detecting anomalies are also introduced. A Monte Carlo simulation study is performed to evaluate the finite signal length performance of the robust estimation and its sensitivity to outliers. The proposed model is applied to a multitemporal normalized difference vegetation index (NDVI) image stack for filtering, prediction, and anomaly detection purposes. The numerical results show the importance of the proposed 3-D-AR model for spatiotemporal remote sensing data interpretation.

Although multimodal remote sensing data analysis can strongly improve the characterization of physical phenomena on Earth's surface, nonidealities and estimation imperfections between records and investigation models can limit its actual information extraction ability. In this article, we aim at predicting the maximum information extraction that can be reached when analyzing a given data set. By means of an asymptotic information theory-based approach, we investigate the reliability and accuracy that can be achieved under optimal conditions for multimodal analysis as a function of data statistics and parameters that characterize the multimodal scenario to be addressed. Our approach leads to the definition of two indices that can be easily computed before the actual processing takes place. Moreover, we report in this article how they can be used for operational use in terms of image selection in order to maximize the robustness of the multimodal analysis, as well as to properly design data collection campaigns for understanding and quantifying physical phenomena. Experimental results show the consistency of our approach.

Extensive attention has been widely paid to enhance the spatial resolution of hyperspectral (HS) images with the aid of multispectral (MS) images in remote sensing. However, the ability in the fusion of HS and MS images remains to be improved, particularly in large-scale scenes, due to the limited acquisition of HS images. Alternatively, we super-resolve MS images in the spectral domain by the means of partially overlapped HS images, yielding a novel and promising topic: spectral superresolution (SSR) of MS imagery. This is challenging and less investigated task due to its high ill-posedness in inverse imaging. To this end, we develop a simple but effective method, called joint sparse and low-rank learning (J-SLoL), to spectrally enhance MS images by jointly learning low-rank HS-MS dictionary pairs from overlapped regions. J-SLoL infers and recovers the unknown HS signals over a larger coverage by sparse coding on the learned dictionary pair. Furthermore, we validate the SSR performance on three HS-MS data sets (two for classification and one for unmixing) in terms of reconstruction, classification, and unmixing by comparing with several existing state-of-the-art baselines, showing the effectiveness and superiority of the proposed J-SLoL algorithm. Furthermore, the codes and data sets will be available at https://github.com/danfenghong/IEEE_TGRS_J-SLoL, contributing to the remote sensing (RS) community.

Hyperspectral unmixing is an important problem for remotely sensed data interpretation. It amounts at estimating the spectral signatures of the pure spectral constituents in the scene (endmembers) and their corresponding subpixel fractional abundances. Although the unmixing problem is inherently nonlinear (due to multiple scattering), the nonlinear unmixing of hyperspectral data has been a very challenging problem. This is because nonlinear models require detailed knowledge about the physical interactions between the sunlight scattered by multiple materials. In turn, bilinear mixture models (BMMs) can reach good accuracy with a relatively simple model for scattering. In this article, we develop a new BMM and a corresponding unsupervised unmixing approach which consists of two main steps. In the first step, a deep autoencoder is used to linearly estimate the endmember signatures and their associated abundance fractions. The second step refines the initial (linear) estimates using a bilinear model, in which another deep autoencoder (with a low-rank assumption) is adapted to model second-order scattering interactions. It should be noted that in our developed BMM model, the two deep autoencoders are trained in a mutually interdependent manner under the multitask learning framework, and the relative reconstruction error is used as the stopping criterion. The effectiveness of the proposed method is evaluated using both synthetic and real hyperspectral data sets. Our experimental results indicate that the proposed approach can reasonably estimate the nature of nonlinear interactions in real scenarios. Compared with other state-of-the-art unmixing algorithms, the proposed approach demonstrates very competitive performance.

The strong urbanization impetus of developing countries leads to various urbanization phenomena such as building constructions, reconstructions and demolitions. It is desirable to monitor and recognize these changes by utilizing temporal and spatial information in an automatic way. In this paper, time-series segmentation and unsupervised classification are combined to mine and recognize segments of temporal change patterns. Our preliminary results show that the proposed approach is effective and reduces the negative consequences due to information inflation.

Although deep learning architectures are nowadays used in several research fields where automatized investigation of large scale datasets is required, the intrinsic mechanisms of deep learning networks are not fully understood yet. In this paper, a new approach for characterizing how information is processed within convolutional neural networks (CNNs) is introduced. Taking advantage of an analysis based on information theory, we are able to derive an index that is associated with the degree of maximum information extraction a CNN can obtain under ideal circumstances as a function of its hyperparameters setup and of the data to be explored. Experimental results on remote sensing datasets show the robustness of our approach. The outcomes of our analysis can be used to optimize the design of CNNs and maximize the information that can be obtained for the considered problem.

The European Space Agency (ESA) Climate Change Initiative (CCI) is aiming, in its current phase, at an accurate description and analysis of land cover (LC) and land cover change (LCC) using high spatial resolution Earth Observation (EO) data. A new high resolution LC map could have a key role in the extraction of the so-called Essential Climate Variables (ECV), and be crucial to understand climate change. Indeed, until now these important variables have been derived by the climate modelling community at the global scale using medium resolution EO data (i.e., with a spatial sampling between 100 and 300 m).

Built-up (BU) area extraction from remote sensing images is important to monitor and manage urbanization and industrialization. In this letter, we propose two BU area extraction techniques based on the analysis of fully polarimetric synthetic aperture radar (PolSAR) data. Both methods exploit the geodesic distance on the unit sphere in the space of Kennaugh matrices. The first method is based on the three dominant scattering types in the scene and compares them with scattering models; if any of them matches with BU type elementary scattering models, then the pixel is said to belong to a BU area. The second method is based on a novel PolSAR BU index (RBUI) composed by considering scattering mechanisms from BU structures. The two proposed techniques are validated on two different urban scenes, one acquired at C-band by RADARSAT-2 and other at L-band by ALOS-2 SAR sensors.

This article introduces the Rayleigh autoregressive moving average (RARMA) model, which is useful to interpret multiple different sets of remotely sensed data, from wind measurements to multitemporal synthetic aperture radar (SAR) sequences. The RARMA model is indeed suitable for continuous, asymmetric, and nonnegative signals observed over time. It describes the mean of Rayleigh-distributed discrete-time signals by a dynamic structure including autoregressive (AR) and moving average (MA) terms, a set of regressors, and a link function. After presenting the conditional likelihood inference for the model parameters and the detection theory, in this article, a Monte Carlo simulation is performed to evaluate the finite signal length performance of the conditional likelihood inferences. Finally, the new model is applied first to sequences of wind speed measurements, and then to a multitemporal SAR image stack for land-use classification purposes. The results in these two test cases illustrate the usefulness of this novel dynamic model for remote sensing data interpretation.

In this paper, we introduce a new approach for investigation of polarimetric Synthetic Aperture Radar (PolSAR) images for oil slick analysis. Our method aims at enhancing discrimination of oil types by exploring the polarimetric features that can be produced by processing PolSAR scenes without dimensionality reduction. Taking advantage of a mixture description of the interactions among classes within the dataset and a characterization of their intra- and inter-class variability, our algorithm is able to quantify the areal coverage of different elements. These estimates can be used to hence improve classification. Experimental results on a PolSAR dataset acquired by unmanned aerial vehicle (UAV) on oil slicks in open water show the capacity of our method.

The data acquired by the Inference Imaging Spectrometer (IIM) sensor on board of the Chinese Chang'E-1 mission can be used to infer important information on the Moon surface composition. In this work, the multi-path and multi-reflection phenomena occurring on its rugged surface recorded at the IIM rather coarse resolution (200m) are described by means of nonlinear spectral analysis based on the p-linear mixture model (pLMM) and the p-harmonic mixture model (pHMM). The analysis by pLMM and pHMM provides details on the materials and elements on the Moon surface, and their abundance distribution and fractional cover can be properly estimated without any a priori information on its chemical composition. Mineral map extractions using pLMM and pHMM have been considered and compared with those obtained by means of the modified partial least squares regression (PLSR) methodology, assessing the reliability and accuracy of the pLMM- and pHMM-based approach.

There is an increasing trend at using social media data to map human activity, to the point that some authors have suggested that these data may be even better than Earth Observation (EO) data to map urban areas. This work introduces a novel approach to map urban areas using SAR images from Sentinel-1 data, and exploiting either Twitter or Weibo data for two different cities, Beijing and Taipei. The use of different social media data results into different urban extent maps, and for different cities the same approach provides different accuracy values for the same area, according to the availability of either social media data set. Instead, urban extents obtained by means of EO data only are more stable. Indeed, the adoption of a specific social platform depends on a number of geographical, economic and political factors, while remote sensing is based on physics.

Nonlinear unmixing of hyperspectral images has been a very challenging research problem, as it needs to consider the physical interactions between the sunlight scattered by multiple materials. In this paper, we propose a new approach for nonlinear unmixing which is based on multi-task learning (MTL) with low-rank matrix factorization (LRMF). The proposed approach establishes two tasks to conduct the unmixing problem under a nonlinear mixing model. In the first task, we employ LRMF to obtain endmember signatures and their corresponding abundance fractions simultaneously. Then, the second task uses LRMF to solve interactions from multiple scattering. The effectiveness of the proposed method is verified by using real hyperspectral data. Compared with other state-of-the-art nonlinear unmixing algorithms, the proposed approach demonstrates very competitive performance.

Nonlinear unmixing algorithms are playing a key role in modern earth observation analysis thanks to their ability to characterize complex phenomena occurring in the instantaneous field of view. When unmixing hyperspectral images according to nonlinear mixture models by means of state-of-the-art methods, actual abundances of the elements in the scene can be only indirectly estimated. Thus, the reliability of the investigation can be dramatically jeopardized, hence degrading the accuracy of the characterization of the surface composition. In order to overcome this issue, we propose in this paper a nonlinear programming scheme that aims at providing direct estimation of the end members fractions. The method we introduce is based on a structural optimization approach where the abundances are directly assessed, so that no epistemic uncertainties are injected in the framework. Experimental results show that the proposed method is able to deliver accurate and reliable estimates of these quantities in hyperspectral images.

Spectral unmixing is a technique for remotely sensed image interpretation that expresses each (possibly mixed) pixel as a combination of pure spectral signatures (endmembers) and their fractional abundances. In this paper, we develop a new technique for unsupervised unmixing which is based on a deep autoencoder network (DAEN). Our newly developed DAEN consists of two parts. The first part of the network adopts stacked autoencoders (SAEs) to learn spectral signatures, so as to generate a good initialization for the unmixing process. In the second part of the network, a variational autoencoder (VAE) is employed to perform blind source separation, aimed at obtaining the endmember signatures and abundance fractions simultaneously. By taking advantage from the SAEs, the robustness of the proposed approach is remarkable as it can unmix data sets with outliers and low signal-to-noise ratio. Moreover, the multihidden layers of the VAE ensure the required constraints (nonnegativity and sum-to-one) when estimating the abundances. The effectiveness of the proposed method is evaluated using both synthetic and real hyperspectral data. When compared with other unmixing methods, the proposed approach demonstrates very competitive performance.

This paper introduces the topic of urban change detection by means of fully and dual polarized multitemporal spaceborne SAR data. The aim is to be able to detect not only urban extent expansion/shrinking phenomena, but changes that may affect both the two- and the three-dimensional landscape. For fully polarized SAR, we further validate and refine a recently introduced approach based on superpixel segmentation followed by the computation of selected polarimetric parameters and the analysis of the difference among feature vectors using specialized distance metrics. As for dual polarized SAR data, a combination of interferometric and polarimetric features within the same framework is proved to perform decently, but no real advantage if using polarimetry is obtained. Test cases using UAVSAR and Sentinel-l data are presented in support of the proposed techniques.

The automatic extraction of building footprints from remotely sensed images has been used for updating geospatial databases in urban areas [1]. The launch of High Resolution Spacebome Stereo (HRSS) sensors (e.g. GeoEye, WorldView, QuickBird) started a new era by providing the possibility to obtain stereo images and 3D maps from space [2]. Indeed, building identification, reconstruction, and change detection have been carried out using stereo image matching, as well as 3D edge matching techniques [3,5-6]. As stated in [3], 3D edge matching based on stereo images delivers promising results, but only if the buildings are large enough with respect to the spatial resolution of the data, have a simple rectangular shape, and a good radiometric contrast compared to surrounding objects. As a matter of fact, although 3D edge matching using very high resolution aerial images can reconstruct building footprints in detail [7], using space-borne images the same approach may encounter issues, particularly where building outlines are not clearly detected in both epipolar images. Additionally, although image matching delivers a DSM representing buildings heights, building size and shapes extracted from this DSM are usually overestimated, so that auxiliary information is required.

This work introduces two feature fusion techniques that exploit previously developed algorithms for urban extent extraction from multispectral and SAR spaceborne data, adapting them to the joint use of Sentinel-l (S1) and Sentinel-2 (S2) data sets. The approaches aim at exploiting the finer spatial and spectral resolution of multispectral S2 data as well as the double bounce backscatter effect that is common to all built-up areas in SAR S1 data. To this aim, we introduce first a simplified and less computational demanding version of the Urban Extractor (UEXT) algorithm, recently introduced for urban extent extraction from S1 data, and improve its results by two different ways of selecting the seed pixels involved in UEXT by means of the urban extent maps extracted from S2 using the normalized difference spectral vector (NDSV), whose application for national and regional extraction of human settlements have already proved as very effective. Experimental results for Rio de Janeiro and Beijing show the improvements obtained by considering one of the two proposed techniques, and explains while the other one fails in achieving similar results.

Accurately estimating the urbanization process is a key-factor for the actual implementation of the sustainable development goals identified by transnational institutions and agencies. In order to retrieve precise characterization of the anthropogenic extents and a sound human-environment interaction assessment, the analysis of Earth observations (EOs) plays a crucial role. Especially, the use of nonlinear spectral investigation can improve the description of geometrically and morphologically complex scenes, so that anthropogenic settlements and dynamics can be properly outlined. In this paper, we propose a novel method for directly assessing the distribution of materials and elements in hyperspectral images by means of a structural optimization approach. Experimental results show how the proposed approach is able to deliver accurate and reliable characterization of urban materials and extents.

Air quality is strongly affecting human lifestyle all over the world, and its impact is apparent on healthcare, sustainable development, welfare and public administration policies. Accurate understanding of the polluting processes requires to analyze huge volumes of records, so that significant patterns and regularities can be detected. In this paper, we introduce a framework to explore the air pollution dynamics over all Europe by means of a data driven feature extraction approach. Taking advantage of MODIS records, we are able to investigate daily trends of air quality from 2003 to 2016. By means of an automatic learning scheme based on mutual information maximization, we extract the most significant patterns in the dataset. Experimental results show that the proposed approach is able to identify relevant air pollution trends that can be associated with specific physical phenomena on ground.

In this paper, we propose a deep auto-encoder network for the unmixing for hyperspectral data with outliers and low signal to noise ratio. The proposed deep auto-encoder network composes of two parts. The first part of the network adopts stacked non-negative sparse auto-encoder to learn the spectral signatures such that to generate a good initialization for the network. In the second part of the network, a variational auto-encoder is employed to perform unmixing, aiming at the endmember signatures and abundance fractions. The effectiveness of the proposed method is verified by using a synthetic data set. In our comparison with other state-of-the-art unmixing methods, the proposed approach demonstrates highly competitive performance.

A framework based on color feature extraction of social media photos and correlation analysis with air quality parameters is proposed to monitor environmental change. More specifically, photos of the Beijing Olympic Park from Panoramio website have been analyzed as a case study. The aerosol optical depth data at 500 nm wavelength (AOD 500) obtained from sun-photometer observation network station has been used as reference. Results show a proof of concept that social media photos have an interesting potential for air pollution estimate and remote sensing parameter validation with a low cost.

Spectral-spatial classification has been widely exploited for hyperspectral imagery. However, current methods either focus on local spatial similarity or global nonlocal self-similarity (NLSS). In this paper, we propose novel methods to couple both global spatial similarity and local spectral similarity together in a single framework. In particular, our approaches exploit global spatial similarity by searching non-overlap nonlocal patches, whereas spectral similarity is determined locally within the found patches. Experimental results on two real hyperspectral data sets demonstrate the efficiency of the proposed methods, with 5%-7% (overall classification accuracy) improvements over approaches that only consider either global or local similarity.

The possibility theory, which is an extension of fuzzy sets and fuzzy logic, has shown considerable potential for solving active learning (AL) problems, particularly for multiclass scenarios’ classification. Hence, two recently proposed fuzzy multiclass AL algorithms (classification ambiguity (CA) and fuzzy C‐order ambiguity (FCOA)) are investigated to properly generalise them for classifying hyperspectral images, and two improved versions of the CA and FCOA are proposed. In addition to comparing the performances of the original and improved algorithms, several other state‐of‐the‐art AL methods are evaluated, such as breaking ties, margin sampling, and multi‐class level uncertainty, with or without diversity criteria such as angle‐based diversity (ABD), clustering‐based diversity (CBD), and enhanced clustering‐based diversity (ECBD). Tests on two benchmark hyperspectral images confirm that the proposed improved algorithms are superior to and more effective than the original ones.

In this letter, an unsupervised 2-D and 3-D urban change detection scheme is proposed exploiting Quad-PolSAR data. Changes are extracted by segmenting the data into superpixels, to enhance the balance among change components and increase estimability of prior distributions. Positive and negative change components for built-up areas, in both the horizontal and the vertical directions, are properly extracted by assuming a multivariate Gaussian mixed model applied to a subset of polarimetric parameters at the superpixel level. The proposed method is tested on multitemporal Quad-PolSAR images and the results confirm its effectiveness. The selection of polarimetric decomposition measures that are most useful to the task is also experimentally justified.

As an unsupervised learning tool, autoencoder has been widely applied in many fields. In this letter, we propose a new robust unmixing algorithm that is based on stacked nonnegative sparse autoencoders (NNSAEs) for hyperspectral data with outliers and low signal-to-noise ratio. The proposed stacked autoencoders network contains two main steps. In the first step, a series of NNSAE is used to detect the outliers in the data. In the second step, a final autoencoder is performed for unmixing to achieve the endmember signatures and abundance fractions. By taking advantage from nonnegative sparse autoencoding, the proposed approach can well tackle problems with outliers and low noise-signal ratio. The effectiveness of the proposed method is evaluated on both synthetic and real hyperspectral data. In comparison with other unmixing methods, the proposed approach demonstrates competitive performance.

In this letter, a new method for higher order nonlinear hyperspectral unmixing is introduced. The proposed scheme relies on the harmonic description of the endmembers contributions to characterize the interactions among the materials showing up in the given scenes. Moreover, it aims at directly estimating the probability of occurrence of each material in the images, so to provide an accurate quantification of the endmembers also in complex scenarios. Experimental results carried out on synthetic and real data sets show that the proposed method is able to obtain good unmixing performance when compared to other state-of-the-art architectures.

Accurately estimating the elements in Earth observations is crucial when assessing specific features such as air quality index, water pollution, or urbanization process behavior. Moreover, physical-chemical composition can be retrieved from hyperspectral images when proper spectral unmixing architectures are employed. Specifically, when linear and nonlinear combinations of endmembers (pure spectral components) are accurately characterized, hyperspectral unmixing plays a key role in understanding and quantifying phenomena occurring over the instantaneous field-of-view. Thus, reliable detection of nonlinear reflectance behavior can play a key role in enhancing hyperspectral unmixing performance. In this paper, two new methods for adaptive design of mixture models for hyperspectral unmixing are introduced. One of the methods relies on exploiting geometrical features of hyperspectral signatures in terms of nonorthogonal projections onto the space induced by the endmembers' spectra. Then, an iterative process aims at understanding the order of local nonlinearity that is displayed by each endmember over every pixel. An improved version of an artificial neural network-based approach for nonlinearity order information is also considered and compared. Experimental results show that the proposed approaches are actually able to retrieve thorough information on the nature of the nonlinear effects over the image, while providing excellent performance in reconstructing the given data sets.

Land cover mapping is usually characterized by a multiscale classification, generally because the image objects are better identified at multiple scales. Consequently, the features are usually extracted at different resolutions, and then classified by means of the same classifier (or a classifier ensemble). Unfortunately, this approach requires to select each time different scales for scenes containing (very) different classes. In these situations, a single multi-scale analysis may not be sufficient. The Hierarchical Binary Decision Tree (HBDT) approach, instead, combines different processing chains (composed by feature selection and classification steps) and automatically adapts to the spatial and spectral properties of the classes to be recognized in a scene. The HBDT algorithm has already proved to be efficient for single data sources. In this paper we propose two different data fusion techniques exploiting HBDT. These two techniques refer one to the fusion of results at the decision level, the other to fusion at the feature level. This work compares the performances of these two techniques using a data set of two city areas, i.e. Pavia (Italy) and Beijing (P.R. China), using HR and VHR optical and SAR data.

Hyperspectral unmixing frameworks are ultimately designed to understand and quantify the actual distribution of endmembers in a given scene. Assessing the percentage of each material is typically cumbersome, especially in images characterized by complex combinations of spectral signatures. In this work, we present a nonlinear programming scheme that aims at providing direct estimation of the endmembers fractions. Experimental results show that the proposed method is able to deliver accurate and reliable estimates of these quantities in hyperspectral images.

Endmember extraction is a fundamental task in spectral unmixing of remotely sensed hyperspectral images. In this work, we develop a new robust algorithm for endmember extraction which is based on a nonnegative sparse autoencoder. The proposed approach is based on two main steps. First, it uses an automatic sampler approach with local outlier factor and affinity propagation to intelligently gather a set of training samples. Then, a set of endmember signatures are extracted from the selected training samples by the nonnegative sparse autoencoder. Taking advantage from both automatic sampling and nonnegative sparse autoencoding, the proposed method can tackle problems with outliers. The effectiveness of the proposed method is verified by using simulated data. In our comparison with other state-of-the-art endmember extraction methods, the proposed approach demonstrates highly competitive performance.

Automatic road extraction from synthetic aperture radar (SAR) imagery has been studied with success in the past two decades. However, a method that combines full interferometric SAR (InSAR) information is as yet missing. In this paper, we present an algorithm toward robust road extraction by fully exploring the multitemporal InSAR covariance matrix. To improve the detection performance and reduce false alarm ratio, intensity and coherence are first accurately estimated without loss of image resolution by homogeneous pixel selection and robust estimators. After the identification of road candidates from each quantity using multiscale line detectors, novel information fusion rules are applied to integrate the extracted results and generate the final road network. The method is tested and quantitatively evaluated on TerraSAR-X data sets depicting two scenes where complex road features make it hard for standard SAR-based methods. The experimental results show that the new method can achieve satisfactory detection performances.

Information theory has recently become an interesting topic in earth observation data management and analysis, since it can provide important information on hidden interactions and correlations among the considered data records. Although several methods have been proposed and implemented to efficiently extract a proper set of features and deliver accurate image investigation, classification, and segmentation, these architectures show drawbacks when the data sets are characterized by complex interactions among the samples. In this paper, a new approach based on information theory for automatic pattern recognition is introduced for accurate classification of remotely sensed data. Experimental results carried out on real data sets show the validity of the proposed approach.

In order to provide reliable information about the instantaneous field of view considered in hyperspectral images through spectral unmixing, understanding the kind of mixture that occurs over each pixel plays a crucial role. In this paper, in order to detect nonlinear mixtures, a method for fast identification of linear mixtures is introduced. The proposed method does not need statistical information and performs an a priori test on the spectral linearity of each pixel. It uses standard least squares optimization to achieve estimates of the likelihood of occurrence of linear combinations of endmembers by taking advantage of the geometrical properties of hyperspectral signatures. Experimental results on both real and synthetic data sets show that the aforesaid algorithm is actually able to deliver a reliable and thorough assessment of the kind of mixtures present in the pixels of the scene.

Multi-scale classification is an important tool for urban image classification, because the objects in an urban scene may have very different spatial scales. In technical literature, this idea is usually performed either using the same classifier (ensemble of classifiers) at multiple resolutions, or working on sets of features at multiple scales. Although this may be very efficient approach for peculiar situations, it is not as effective, as it does not adapt to different classes. Accordingly, it can be useful to design a procedure based on the possibility to assemble specialized (multi-scale) classification chains automatically, according to the classes to be recognized and their peculiar scales or features. We present a methodology that combines different processing chains (each one composed by a feature selection and a classification step) and automatically adapts to the properties of the classes available in an urban scene.

Mapping and monitoring urban area extents is a very relevant task for many applications related to risks, economic activities, population mapping and the interaction between people and the natural environment. This task can be accomplished using VHR and HR optical data. The main drawback of using the optical data is the irregular presence of the clouds, making mapping by these sensors completely useless in many cases. Even when the clouds are small and irregular, the extracted urban area extents may be erroneous and/or incomplete, and many uncertainties show up in thematic maps. In this paper, we prove that the joint use of algorithms for urban extent extraction, cloud detection and masking, and image reconstruction provide a clear processing chain able to overcome this issue and to increasing the reliability of the extracted urban area extents.

In order to provide reliable information on the instantaneous field-of-view considered in hyperspectral images through spectral unmixing, understanding the kind of mixture that occurs over each pixel plays a crucial role. In this paper, a new method for fast detection of linear mixtures is introduced. The proposed method does not need statistical information and performs an a priori test on the spectral linearity of each pixel. It uses standard least squares optimization to achieve estimates of the likelihood of occurrence of linear combinations of endmembers by taking advantage of geometrical properties of hyperspectral signatures. Experimental results on synthetic datasets show how the aforesaid algorithm is actually able to deliver a reliable and thorough assessment of the kind of mix on the scene.

Active learning (AL) methods that select unlabeled samples only querying by informative measures (i.e., uncertainty and/or diversity criteria) have been extensively investigated. However, these methods usually do not exploit the manifold structure of the unlabeled data from the geometrical point of view, a choice that might lead to a sample bias and consequently undesirable performances. To control and possibly overcome such drawbacks, this paper explores AL methods based on joint informative and manifold structure representative sampling (JI-MSRS). In JI-MSRS, a portion of the unlabeled samples that are added at each iteration is selected according to the informative measures, whereas another portion is selected according to their capability to represent the data cluster structure. Four popular manifold learning methods, namely, principle component analysis (PCA), linear discriminant analysis, kernel PCA, and neighborhood preserving embedding, are used to model the data structure. Then, Delaunay triangulation nets are used to build a discrete approximation of the geometrical structure of the unlabeled data cloud in a low-dimensional space. To show the effectiveness of this novel sampling strategy, results on three real multi-/hyperspectral data sets are presented, adding a thorough comparison with other state-of-the-art AL techniques. In comparison to conventional AL heuristics, the proposed techniques are able to obtain competitive or even better classification accuracy values.

Advances in remote sensors are providing exceptional quantities of large-scale data with increasing spatial, spectral and temporal resolutions, raising new challenges in its analysis, e.g. those presents in classification processes. This work presents the architecture of the InterIMAGE Cloud Platform (ICP): Data Mining Package; a tool able to perform supervised classification procedures on huge amounts of data, on a distributed infrastructure. The architecture is implemented on top of the MapReduce framework. The tool has four classification algorithms implemented taken from WEKA's machine learning library, namely: Decision Trees, Naïve Bayes, Random Forest and Support Vector Machines. The SVM classifier was applied on datasets of different sizes (2 GB, 4 GB and 10 GB) for different cluster configurations (5, 10, 20, 50 nodes). The results show the tool as a potential approach to parallelize classification processes on big data.

Image segmentation is a critical step in image analysis, and usually involves a high computational cost, especially when dealing with large volumes of data. Given the significant increase in the spatial, spectral and temporal resolutions of remote sensing imagery in the last years, current sequential and parallel solutions fail to deliver the expected performance and scalability. This work proposes a scalable and efficient segmentation method, capable of handling efficiently very large high resolution images. The proposed solution is based on the MapReduce model, which offers a highly scalable and reliable framework for storing and processing massive data in cloud computing environments. The solution was implemented and validated using the Hadoop platform. Experimental results attest the viability of performing region growing segmentation in the MapReduce framework.

Satellite acquisitions from LANDSAT (LS) and CBERS programs are widely used in monitoring land cover dynamics. In the acquired products, clouds form opaque objects are obscuring parts of the scene and preventing a reliable extraction of information from these areas. Consequently, cloud shadows create similar problems, as the reflected intensity of the shadowed areas is highly reduced, generating additional info gaps. The problem can be handled by replacing clouds/shadows pixels from other close-date acquisitions, but that would assume a prior knowledge of the spatial distribution of clouds and their corresponding shadows in a scene. This research introduces a method that provides the clouds/shadows layers and their percentage in LS (TM & ETM+) and CBERS (HRCC) scenes. The approach relies on a set of literature indicators to create a composite image that enhances the visual differentiation of clouds/shadows from other objects. The created composite RGB are then warped to a relative luminance raster calculated from the linear bands components. Afterwards, the raster is processed by a K-means unsupervised classifier with a definite number of classes in order to isolate the target-layer pixels. Next, the statistical mode for the population of each class is calculated, compared and used to select the cloud/shadow class automatically, and finally the results are refined by a set of morphological filters. The processing chain avoids the usage of thresholds and highly reduces the user intervention. The achieved outcomes on various test cases are promising and stable, and encourage further developments.

Recent developments in data acquisition, storage, mining and maintenance have allowed the flourishing of several multi-disciplinary research fields, which can be stated, defined and carried out according to the so-called Big Data paradigm. In this environment, the investigation and analysis of interactions between human phenomena and natural events play a key-role, as they can be fundamental for several applications, from sustainable development to community policy design and short-, medium- and long-range resource allocation planning. In this paper, we provide a study of the interplay between air pollution (as estimated by remotely sensed data processing) and clinical records, so that inferences and correlations among black particulate concentration, micro- and macro-vascular disease onsets and hospitalization tracks can be efficiently drawn. We focused on the second order administrative area of the city of Pavia, Italy, on 2013. Experimental results show how effective connections between the estimated air quality and the hospitalizations behavior can be accurately drawn and derived.

As remotely sensed Big Data applications in astrophysics research have been flourishing in the last decade, the need for a new class of techniques and methods for efficient storage, compression, retrieval and investigation of astronomical datasets has become urgent. In this paper, a novel strategy for lossless compression of large datasets composed by remote sensing records is introduced. Specifically, the new approach aims at describing each sample of the given dataset as a point living within a convex hull in a multidimensional space. Thus, the proposed framework aims at characterizing every sample as a nonlinear combination of the extremal points of the aforesaid multidimensional simplex. Therefore, efficient compression can be achieved by describing those samples by the parameters that drive the nonlinear mixture only. Experimental results show how the proposed architecture can effectively deliver great compression performance for both Earth observations and planetary records.

Remote sensing is one of the most common ways to extract relevant information about Earth and our environment. Remote sensing acquisitions can be done by both active (synthetic aperture radar, LiDAR) and passive (optical and thermal range, multispectral and hyperspectral) devices. According to the sensor, a variety of information about the Earth's surface can be obtained. The data acquired by these sensors can provide information about the structure (optical, synthetic aperture radar), elevation (LiDAR), and material content (multispectral and hyperspectral) of the objects in the image. Once considered together their complementarity can be helpful for characterizing land use (urban analysis, precision agriculture), damage detection (e.g., in natural disasters such as floods, hurricanes, earthquakes, oil spills in seas), and give insights to potential exploitation of resources (oil fields, minerals). In addition, repeated acquisitions of a scene at different times allows one to monitor natural resources and environmental variables (vegetation phenology, snow cover), anthropological effects (urban sprawl, deforestation), climate changes (desertification, coastal erosion), among others. In this paper, we sketch the current opportunities and challenges related to the exploitation of multimodal data for Earth observation. This is done by leveraging the outcomes of the data fusion contests, organized by the IEEE Geoscience and Remote Sensing Society since 2006. We will report on the outcomes of these contests, presenting the multimodal sets of data made available to the community each year, the targeted applications, and an analysis of the submitted methods and results: How was multimodality considered and integrated in the processing chain? What were the improvements/new opportunities offered by the fusion? What were the objectives to be addressed and the reported solutions? And from this, what will be the next challenges?

Classification and spectral unmixing are two important techniques for hyperspectral data exploitation. Traditionally, these techniques have been exploited independently. In this paper, we propose a new technique that exploits their complementarity. Specifically, we develop a new framework for semisupervised hyperspectral image classification that naturally integrates the information provided by discriminative classification and spectral unmixing. The idea is to assign more confidence to the information provided by discriminative classification for those pixels that can be easily catalogued due to their spectral purity. For those pixels that are more highly mixed in nature, we assign more confidence to the information provided by spectral unmixing. In this case, we use a traditional spectral unmixing chain to produce the abundance fractions of the pure signatures (endmembers) that model the mixture information at a subpixel level. The decision on which source of information is prioritized in the process is taken adaptively, when new unlabeled samples are selected and included in our semisupervised framework. In this regard, the proposed approach can adaptively integrate these two sources of information without the need to establish any weight parameters, thus exploiting the complementarity of classification and unmixing and selecting the most appropriate source of information in each case. In order to test our concept, which has similar computational complexity as traditional semisupervised classification strategies, we have used two different hyperspectral data sets with different characteristics and spatial resolution. In our experiments, we consider two different discriminative classifiers: multinomial logistic regression and probabilistic support vector machine. The obtained results indicate that the proposed approach, which jointly exploits the features provided by classification and spectral unmixing in adaptive fashion, offers an effective solution to improve classification performance in hyperspectral scenes containing mixed pixels.

Abstract—Hyperspectral image classification has been an active topic of research in recent years. In the past, many different types of features have been extracted (using both linear and nonlinear strategies) for classification problems. On the one hand, some approaches have exploited the original spectral information or other features linearly derived from such information in order to have classes which are linearly separable. On the other hand, other techniques have exploited features obtained through nonlinear transformations intended to reduce data dimensionality, to better model the inherent nonlinearity of the original data (e.g., kernels) or to adequately exploit the spatial information contained in the scene (e.g., using morphological analysis). Special attention has been given to techniques able to exploit a single kind of features, such as composite kernel learning or multiple kernel learning, developed in order to deal with multiple kernels. However, few

In this paper, two different approaches are proposed for the estimation of building footprints and number of stories using high resolution space-borne images. To this aim, semiglobal matching (SGM) is used to generate normalized digital surface models (nDSM) from stereo pairs. Alternatively, a height-from-shadow approach (called “shadow-raiser”) is implemented by detecting building rooftops and related shadow regions. Using associated lengths of shadow, the building heights are computed based on sun elevation and azimuth. The results of the proposed algorithms using IKONOS and GeoEye images demonstrate promising results with SGM, although the building dimensions are usually overestimated. In contrast, shadow-raiser delivers good results only if the building-shadow pair is correctly detected. Moreover, it suffers from an overestimation for building height if shadow areas are mixed up with occluded areas, vegetation or roads.

This paper presents a new multiple feature learning approach for accurate spectral-spatial classification of hyperspec-tral images. The proposed method integrates multiple features based on the logarithmic opinion pool. We consider subspace multinomial logistic regression for classification as it exhibits a flexible structure for the combination of multiple features through the posterior probability. At the same time, it is able to cope with highly mixed hyperspectral data and with the presence of limited training samples. In this work, we considered lowpass filtering and morphological attribute profiles for spatial feature extraction. Our experimental results with a real hyperspectral images collected by the NASA Jet Propulsion Laboratory's Airborne Visible Infra-Red Imaging Spectrometer (AVIRIS) indicate that the proposed method exhibits state-of-the-art classification performance.

This work is devoted to the fusion of spaceborne SAR urban extent extractions at multiple spatial resolutions. The results of already existing urban settlement detection algorithms are fused in order to enhance the accuracy and to increase the reliability of the overall urban extent maps. Simple, but reasonable logical combinations are used as fusion operators. In order to find out if the fusion at the decision level leads to better results, the accuracies of both, the single image as well as the fused results, are assessed and compared. Results show that the fusion exhibits lower Commission errors and slightly higher Omission errors, resulting in a more reliable urban area extraction.

This paper describes a methodology to extract a consistent human settlement extent layer using Landsat data and its implementation in the Google Earth Engine platform. The approach allows the extraction of the human extents by means of the existing Landsat 5 and 7 data sets, allowing a multitemporal analysis of the evolution of human settlements at 30 m spatial resolution. Since human settlements are the main proxy to people geographical distribution, this layer may serve as a mean to disaggregate people counts in multiple time instants, with a consistent accuracy along more than 20 years. The approach is tested against available global data sets and existing ground truth data at the same spatial resolution, as well as with extents manually extracted from VHR data for large urban areas in Brazil.

In cognitive radio networks (CRNs), knowledge of the primary users (PUs) position can be used to avoid harmful interference to the primary network, while at the same time be exploited to improve CR performance. In this paper, a localization algorithm is developed to calculate PUs position and a novel location based CR (LCR) routing protocol is proposed that has the following properties: (i) it considers the existence of heterogeneous PUs, (ii) exploits PUs location information, (iii) jointly selects spectrum and route, (iv) protects PUs from interference. Clusters of CRs are defined according to the spectral characteristics in a given location area, and the LCR routing protocol acts in two steps: intra-cluster and inter-cluster. Simulations are conducted in terms of CR end-to-end performance and PUs collision risk. Results reveal the importance of formulating routing protocol in terms of PU protection, which is a unique features in CR networks.

The characterization of urban areas can be improved considerably by combining spectral and spatial features. As a matter of fact, depending on objects of interest in a specific application, the exploitation of both types of features at multiple spatial resolutions is required. This paper proposes a decision fusion method that relies on both spectral and textural features. The proposed approach is able to produce different classification results based on distinct partitions of the same input data set. Experiments conducted on CBERS-2B data demonstrate a significant performance improvement brought by the combination of spectral and textural features in comparison to the use of only spectral features to describe the image objects.

The availability of fine spatial resolution earth observation (EO) data has been always considered as a plus for human settlement monitoring from space. It calls, however, for new and efficient data processing tools, capable to manage huge data amounts and provide information for urban area monitoring, management, and protection. Issues related to data processing at the global level imply multiple-scale processing and a focused data interpretation approach, starting from human settlement delineation using spatial information and detailing down to material identification and object characterization using multispectral and/or hyperspectral data sets. This paper attempts to provide a consistent framework for information processing in urban remote sensing, stressing the need for a global approach able to exploit the detailed information available from EO data sets.

In order to compare fusion algorithms considering both active and passive remotely sensed data, a few well-known techniques, including Brovey, Gram-Schmidt spectral sharpening (GS), Hue Saturation Value (HSV), Principal Component Analysis spectral sharpening (PCA), and à trous wavelet transform applied in the Hue Intensity Saturation space (ATWT+HIS) are compared with the simple joint analysis of the original SAR and optical images. Experiments are performed using pairs of ALOS ANVIR-2 and PALSAR, SPOT and PALSAR, Landsat TM and ERS data. In this paper, the above mentioned methods and dataset combinations are tested and compared by means of quantitative indexes such as entropy, average gradient (AG), correlation coefficient (CC), deviation index (DI) and classification accuracy. The results obtained demonstrate that classification accuracy values can be improved by using these approaches by as much as 10% with respect to the best achievable value using only optical and SAR data separately. By means of a detailed analysis of the relationship between classification accuracy and quantitative indexes usually considered to evaluate the value of the fused products, our experiments show that larger deviation index (DI) and smaller correlation coefficient (CC) values are usually connected to more accurate classification results.

In this paper we propose two novel methodologies to incorporate spatial information into ensemble classification systems to process hyperspectral data acquired over urban environment. We introduce the methodologies for extending Hierarchical Binary Decision Tree Classification structure based ensemble (HBDTC) and Class probability Membership value based Ensemble (PMVE) structures with capability to use information from the spatial domain while optimizing the classification structure. In current study a Canny edge detector based clustering and region growing based image segmentation are combined to obtain image object features and after optimizing the ensemble structures in the spectral domain a further optimization is carried out using the identified image objects and refinement in the labelling is done. The obtained classification results show great potential to use spectral-spatial ensemble classification structures for generic mapping of the urban environment. In the paper we demonstrate on two different scenes that both HBDTC spatial algorithm and PMVE spatial algorithms outperform ensemble classification without spatial extension, even if coupled with spatial post.

This papers presents a first attempt to analyze and fuse human settlement extents extracted from optical and radar data. Due to the different possible definition of human settlements (artificial land covers, urban land use classes, built-up areas,...) the approaches that are discussed in these paper present different results on the same area. While comparison with ground truth is always feasible, a combination of extraction results from SAR and optical is acceptable only when the same underlying definition of human settlement is considered.

In the framework of the monitoring of structures and infrastructures from environmental disasters, the COSMO-SkyMed constellation has a huge potential, thanks to up to metric spatial resolution, short revisit time, and the day/night all-weather acquisition capability ensured by SAR. This paper focuses on the scientific results of the project “Development and validation of multitemporal image analysis methodologies for multirisk monitoring of critical structures and infrastructures,” funded by the Italian Space Agency. Several change-detection, data-fusion, and feature-extraction techniques, which were developed and experimentally validated in the project for COSMO-SkyMed imagery and for their integration with other data sources (including very high resolution optical data), are described and examples of processing results are discussed.

In this paper, a survey of the techniques and data sets used to evaluate earthquake damages using remote sensing data is presented. After a few preliminary definitions about earthquake damage, their evaluation scale, and the difference between identification of damage “extent” and identification of damage “level,” the advantages and limits of different remote sensing data sets are presented. Furthermore, a survey of proposed algorithms for data interpretation and earthquake damage extraction is presented, and two examples of these algorithms and their results are discussed. According to the outcome of this survey, some open issues are finally presented and discussed, identifying possible research lines as well as working solutions.

The work presented in this paper is devoted to the analysis of SAR images in order to produce at first a human settlements map, followed by a refined classification of the same dataset in order to extract a Land Use Land Cover (LU/LC) map based on CORINE nomenclature. The urban extents are computed using an approach based on Local Indicators of Spatial Association and textural features while the LU/LC map is obtained using a segmentation technique in order to exploit the statical behaviours of areas belonging to the same class. In particular, in this paper the joint use of SAR (for classification) and optical images (for segmentation) on the same area is investigated, together with the comparison of three different segmentation techniques based on different algorithms and approaches. Conducted tests on the area of Shanghai demonstrated that the use of even dated optical images for the segmentation phase increases the accuracy and reveals the potential of the entire processing chain for urban areas monitoring.

Low spatial resolution is a major limitation for remote sensing classification, especially in a urban environment. In this work, we will focus on the simulation of urban area environment at a low spatial resolution, comparable to the new hyperspectral sensors that will be launched in the next few years. The aim is to better understand the possibility offered by the new sensors, in a challenging scenario like the one represented by a highly mixed image. Particular attention is placed on the characteristics of the sensor EnMap, produced by DLR. The experiments conducted on a real data set confirm the challenges posed by low spatial resolution when analyzing a urban environment.

In this paper we present different fusion methodologies able to exploit road extraction from an airborne dual-band SAR acquisition. All the different approaches discussed in this paper profit from the complementary backscattering characteristics of X-band and P-band radars. As a result of our research, it is shown, by comparing road network extraction using different processing chains, that fusion at the information level is the best way to combine the extraction from multi-frequency SAR data. The level of information fusion does however play a role. Specifically, the fusion of road candidates before network regularization provides the most accurate road network extraction in both the considered test areas. Additionally, the research presented in this paper highlights the differences between urban and rural environments, stressing the greater importance of dual-based acquisition outside human settlements.

A method for change detection between two hyperspectral datasets is presented. The iteratively reweighted multivariate alteration detection (IR-MAD) method is used for change detection. The strong changes are first eliminated based on the principal component analysis (PCA) of the difference image and IR-MAD is applied on the datasets after feature reduction with the PCA of the original bands. The method is demonstrated on a bitemporal hyperspectral dataset. The results show good correlation with ground truth.

In this letter, we introduce a methodology to combine decisions of multiple hyperspectral data processing chains using an already tested preselection step and a novel algorithm for the data labeling procedure. More specifically, we exploit a hierarchical binary decision tree (HBDT) optimization algorithm to select the most suitable processing chains for a given mapping problem. Then, a new methodology for decision fusion is introduced, based on weighting the class probability membership values. Experimental results in two test areas show great potentials for the novel procedure, identified as particularly useful for generic mapping of complex environments due to its flexibility and robustness. Moreover, accuracy values are improved with respect to those obtained by HBDT alone.

Supervised classification of hyperspectral images is a very challenging task due to the generally unfavorable ratio between the number of spectral bands and the number of training samples available a priori, which results in the Hughes phenomenon. For this purpose, several feature extraction methods have been investigated in order to reduce the dimensionality of the data to the right subspace without significant loss of the original information that allows for the separation of classes. In this letter, we explore the use of spectral unmixing for feature extraction prior to supervised classification of hyperspectral data using support vector machines. The proposed feature extraction strategy has been implemented in the form of four different unmixing chains and evaluated using two different scenes collected by National Aeronautics and Space Administration Jet Propulsion Laboratory's Airborne Visible/Infrared Imaging Spectrometer. The experiments suggest competitive results but also show that the definition of the unmixing chains plays an important role in the final classification accuracy. Moreover, differently from most feature extraction techniques available in the literature, the features obtained using linear spectral unmixing are potentially easier to interpret due to their physical meaning.

This letter examines the effect of the prior elimination of strong changes on the results of change detection in bitemporal multispectral images using the previously published iteratively reweighted multivariate alteration detection (IR-MAD) method. An initial change mask is calculated by identifying strong changes between two images. By using the mask and hence eliminating the strong changes from the analysis, the IR-MAD method is able to identify a better no-change background. This effect is demonstrated on a multitemporal Landsat Enhanced Thematic Mapper Plus data set from an agricultural region in Germany with substantial improvement in the results even for the scenes which have a large number of changes.

In many practical applications of hyperspectral remotely sensed data, maps of different land cover classes or features of interest are best achieved by means of different processing algorithms and techniques. In this paper, we introduce a novel methodology to build a multistage hierarchical data processing approach that is able to combine the advantages of different processing chains, which may be best suited for specific classes, or simply already available to the data interpreters. The combination process is carried out using a hierarchical hybrid decision tree architecture where, at each node, the most useful input information source, i.e., the processing chain, is used. The structure of the tree is created by using the predicted accuracy level of the whole structure estimated on a validation set. The final maps are achieved by applying the designed framework to the whole data set. The usefulness of the procedure is proved by two instances of a specific application, i.e., vegetation mapping, in mountainous and plain areas.

The IEEE Geoscience and Remote Sensing Society celebrated the 30th International Geoscience and Remote Sensing Symposium (IGARSS 2010) on July 25–30 at the Hilton Hawaiian Village in Honolulu, HI. This article discusses the highlights of IGARSS 2010 and the importance of crowd-sourcing or volunteerd geographic information for the future of remote sensing.

In this paper we introduce a methodology to combine decisions of multiple data processing chains using novel algorithms for the selection of the processing chains to be used and also for the data labeling procedure. More specifically we recall how a Hierarchical Binary Decision Tree designing and optimization algorithm can be used to select the most suitable processing chains, given the dataset and the training and validation data. Then, we introduce a new methodology for the decision fusion of these processing chains by using class probability membership values. The test results show great potential of the introduced methodology, identified as particularly useful for generic mapping of vegetation because of its flexibility and robustness. The latter addition improves the already high accuracy level obtained by Hierarchical Binary Decision on the AVIRIS Indian Pine 1992 dataset. While this improvement is not dramatic in terms of overall accuracy, it is shown that the method is more robust in case of classes that are difficult to discriminate using other techniques.

Automated earthquake damage assessment from post-event only remotely sensed data is highly desirable, especially when new generation, Very High Resolution (VHR) spaceborne data is concerned, lacking extensive pre-event archives. Though, most damage assessment method either rely on human interpretation or on pre-post-event comparison. In this paper we illustrate some possible tracks for investigating damage assessment on post-event only data, focusing on the 6thApril 2009 Abruzzi, Italy earthquake and on related COSMO/SkyMed acquisitions.

It is an honor and pleasure to present the 2010 IGARSS technical program. This year marks the 30th anniversary of IGARSS dating from the 1981 meeting in Washington DC. Entering its fourth decade, IGARSS continues to be the premier conference in remote sensing providing a unique opportunity for the world's experts in related disciplines to interact and advance the state of the art.

A method for automatic identification of changes using regression with neural networks is presented. The regression is iteratively performed by updating the weights of the pixels. The method is applied to a small subset of two Landsat images and the results indicate that the proposed method produces good results.

In this paper, we propose to combine two road extractors from very high resolution synthetic aperture radar scenes: one more successful in rural areas and one explicitly designed for urban areas. In order to get the best combination of both, a rapid mapping filter for discriminating rural and urban scenes is utilized. Finally, the results are fused on a feature level and connected by means of a network optimization. The approach is tested and evaluated on TerraSAR-X data containing complex urban areas and urban-rural fringe scenes.

In this paper we introduce the first steps towards a comprehensive methodology for hyperspectral data analysis suitable for generic vegetation mapping applications. As hyperspectral data is characterized by the large number of narrow wavebands it is challenging to find an optimal solution for data classification. In case of vegetation mapping is even more complicated as the signatures of vegetation classes are not constant over time and space. In this study we investigate a series of different data processing chains for vegetation interpretation and we introduce a novel methodology to build a multistage, hierarchical data processing approach that is able to combine the advantages of different processing chains already available to the users. The data classification is carried out using a binary decision tree structure where at each node the most useful input source is used and the structure of the tree is created by using the predicted accuracy level of the whole structure estimated from test classification of data subsets.

The Sichuan Earthquake on the 12th of May 2008, and the extensive rescue operations following this tragic event, proved the value of high-resolution optical and radar remote sensing during the emergency response. Optical data provide a fast and simple way to value ¿at glance¿ damages while radar sensors can deliver images independent of weather conditions, day and night, and thus in principle can represent a mean to obtain a damage map in the immediate aftermath of an event, providing precious information for intervention planning. On the other hand, SAR data is far more difficult to interpret than optical data both to the expert and non-expert. In this paper we present a case study of damage assessment on the Sichuan earthquake experimenting the use of very high resolution data from both worlds, discussing preliminary results and perspectives.

The Hyperspectral Imaging Network (HYPER-I-NET) is a four-year Marie Curie Research Training Network project designed to build an interdisciplinary European research community focusing on hyperspectral imaging activities. The network is currently formed by a multidisciplinary team composed of fifteen highly experienced European partner organizations. In this paper, we outline the activities that have been carried out in the four main areas covered by HYPER-I-NET: 1) hyperspectral sensor specification, 2) processing chain definition and implementation, 3) calibration, validation and definition of standardization mechanisms, and 4) science applications. Along with the description of the progress made in the four main areas listed above, this paper also describes some of the training and transfer of knowledge activities carried out during the first two years of the project.

This paper presents two methods to fuse SAR and optical data for urban extent extraction. The two methodologies build over single sensor's procedure in order to improve the efficiency of the characterization of the urban environment when more data is available. Results over Pavia and Al Fashir conform the effectiveness of the proposed procedures.

The 2008 Data Fusion Contest organized by the IEEE Geoscience and Remote Sensing Data Fusion Technical Committee deals with the classification of high-resolution hyperspectral data from an urban area. Unlike in the previous issues of the contest, the goal was not only to identify the best algorithm but also to provide a collaborative effort: The decision fusion of the best individual algorithms was aiming at further improving the classification performances, and the best algorithms were ranked according to their relative contribution to the decision fusion. This paper presents the five awarded algorithms and the conclusions of the contest, stressing the importance of decision fusion, dimension reduction, and supervised classification methods, such as neural networks and support vector machines.

The results shown in this paper highlights the usefulness of a recently proposed index to extract hints of built-up areas in remotely sensed images. The novelty of this work is in the application of the approach to a very different data set than the one for which the index was originally developed, i.e. SAR instead of optical data. Due to the different approaches (active vs. passive sensors), wavelengths (optical vs. microwave) and distortion/noise effects (additive vs. speckle noise), it is valuable to find out the advantages and limits of the index results on these new data sets. Moreover, due to the different geometry of acquisition for radar sensors, two different implementations of the same index are considered and compared, adding insights on the suitability of slant-range vs. ground-range analysis of SAR data for built-up area recognition.

This work is devoted to the definition and application of a processing chain for rapid mapping using TerraSAR-X data, The approach is based on a quick extraction of spatial features such as linear and textural elements of the scene, and their combination with the original SAR data. The suitability of such procedure for an operative use is proved by the results shown on a simulated and two real TerraSAR-X data sets.

In this paper, we describe a first approximation to the relevant issue of defining a part of the hyperspectral processing chain in a flexible manner. An ultimate goal of our study is to objectively quantify the impact of different (standard and new) processing stages on the generation of a realistic, user-oriented product in the context of an urban land cover mapping problem by means of hyperspectral data, selected in this work as an application case study for demonstration purposes. Although the proposed study is linked to a specific application domain, our experimental results reveal interesting considerations that may help image analysts in defining customized processing chains based on parameters which can be identified and objectively evaluated a priori, such as available sensor resolution or ancillary information. In addition, our study also demonstrates the importance of incorporating information related to both the spatial and the spectral domain in the different steps that comprise the hyperspectral processing chain; particularly when such chain can take advantage of the combined use of both sources of information as it is the case in the considered urban characterization application.

In this paper a novel method, based on autocorrelation indexes and gray-level co-occurence matrix, for the extraction of urban areas in high resolution SAR images, is presented. It strongly reduces human interpreters' intervention thanks to a high degree of automation within the processing chain and allows a fast and accurate generation of built-up area maps, which can be employed for land mapping and support in relief operations.

The subject of this work is the model selection of kernels with multiple parameters for support vector machines (SVM), with the purpose of classifying hyperspectral remote sensing data. During the training process, the kernel parameters need to be tuned properly. In this work a gradient descent based algorithm is used to estimate the parameters. The selection of multiple parameters is addressed, and an approach based on the analysis of the variance values of individual bands was proposed. Several state of the art kernels were tested. Experiments were conducted on real hyperspectral data. Results obtained with the different approaches/kernels were compared statistically, and showed good results in terms classification accuracies and processing time.

The 2007 Data Fusion Contest that was organized by the IEEE Geoscience and Remote Sensing Data Fusion Technical Committee was dealing with the extraction of a land use/land cover maps in and around an urban area, exploiting multitemporal and multisource coarse-resolution data sets. In particular, synthetic aperture radar and optical data from satellite sensors were considered. Excellent indicators for mapping accuracy were obtained by the top teams. The best algorithm is based on a neural classification enhanced by preprocessing and postprocessing steps.

This paper presents a novel iterative symbol timing recovery (STR) scheme for short burst transmission formats, a paradigm commonly found in modern wireless systems, like, for instance, time division multiple access (TDMA) schemes and future wireless packet data networks. Both data-aided (DA) and decision-directed (DD) solutions are considered and performance is pursued by means of an iterative burst-by-burst scheme which exploits the Farrow structure for the polynomial interpolation filter. The convergence of the algorithm is discussed according to the expectation maximization (EM) framework. Performance is evaluated by simulating 4-QAM and 16-QAM transceivers and simulations results are compared under different modulation orders and channel conditions, for both the decision-directed and data-aided cases.

The 19 papers in this special issue focus on data fusion. The papers are summarized here.

This paper explains some of the the goals and objectives of the newly started HYPER-I-NET Marie Curie Research and Training Network. In particular, the requirements related to the definition and implementation of an efficient, adequate and sufficiently general data processing chain for hyperspectral data analysis are considered. Some of the research lines that are expected to play a central role in the activities of this network are also presented and briefly discussed.

Abstract—This paper addresses the main goals and objec-tives of the Hyperspectral Imaging Network (HYPER-I-NET), a recently started Marie Curie Research Training Network. The project is designed to build an interdisciplinary research community focusing on hyperspectral imaging activities. The core strategy of the network is to create a powerful interdisciplinary synergy between different domains of expertise closely related to hyperspectral imaging activities in Europe, ranging from sensor design and flight operation to data collection, processing, interpretation, and dissemination. Our main goals in this paper are to present the project to the Geoscience and Remote Sensing community and to provide an overview of the planned activities in each sub-activity covered by the network.

Urban area classification of very high resolution optical images relies on the one hand on the precise characterization of homogenous spectral responses within objects. On the other hand, sharp edges between the same objects, usual in man-made environments, have to be correctly detected. These two conflicting requirements make adaptive algorithms more suitable fo the task. The present work is devoted to introduce and validate one of these adaptive algorithms, based on Markov random fields (MRF) and neural networks, the approach works in a separate way on the two parts of the image, homogeneous and non.homogeneous ones, and allows to take into account their peculiarities. As such, it proves to be more reliable and accurate than basic maximum likelihood or even MRF and neural network classifiers considered alone.

This paper presents a joint symbol timing recovery and equalization scheme for short burst transmission formats, useful in modern wireless systems, like, for instance, time- division multiple access schemes and future wireless packet data networks. Both data aided and decision directed solutions are considered and the joint optimization performance is pursued by means of an iterative scheme which exploits a timing error function sampled at symbol rate, implemented by means of the Farrow structure for the interpolation filter. Equalization is obtained by zero forcing linear filtering, considering a baud spaced implementation. Performance is evaluated by simulating QPSK transceivers and simulations results are compared with the ideal solutions for both symbol timing recovery and zero forcing channel equalization under frequency selective multipath fading.

In January 2006, the Data Fusion Committee of the IEEE Geoscience and Remote Sensing Society launched a public contest for pansharpening algorithms, which aimed to identify the ones that perform best. Seven research groups worldwide participated in the contest, testing eight algorithms following different philosophies [component substitution, multiresolution analysis (MRA), detail injection, etc.]. Several complete data sets from two different sensors, namely, QuickBird and simulated Pleiades, were delivered to all participants. The fusion results were collected and evaluated, both visually and objectively. Quantitative results of pansharpening were possible owing to the availability of reference originals obtained either by simulating the data collected from the satellite sensor by means of higher resolution data from an airborne platform, in the case of the Pleiades data, or by first degrading all the available data to a coarser resolution and saving the original as the reference, in the case of the QuickBird data. The evaluation results were presented during the special session on data fusion at the 2006 international geoscience and remote sensing symposium in Denver, and these are discussed in further detail in this paper. Two algorithms outperform all the others, the visual analysis being confirmed by the quantitative evaluation. These two methods share the same philosophy: they basically rely on MRA and employ adaptive models for the injection of high-pass details.

The 19 articles in this special issue focus on remote sensing for major disaster prevention, monitoring, and assessment. Topics include earthquakes and landslides, tsunami, hurricanes and typhoons, floods and fires, as well as papers with a broader focus, highlighting innovative tools and procedures to exploit Earth observation data.

In this paper, a mapping procedure exploiting object boundaries in very high-resolution (VHR) images is proposed. After discrimination between boundary and nonboundary pixel sets, each of the two sets is separately classified. The former are labeled using a neural network (NN), and the shape of the pixel set is finely tuned by enforcing a few geometrical constraints, while the latter are classified using an adaptive Markov random field (MRF) model. The two mapping outputs are finally combined through a decision fusion process. Experimental results on hyperspectral and satellite VHR imagery show the superior performance of this method over conventional NN and MRF classifiers.

In this paper, the problem of rapid earthquake damage detection in urban areas using multitemporal synthetic aperture radar data is addressed. It is shown that the combination of intensity and phase features enhances the damage pattern extracted from the data temporal stack using a spatially aware classifier. Moreover, the use of ancillary data, easily available for urban areas, further improves the accuracy by discarding uninteresting parts of the scene and forcing homogeneous classification within city blocks to avoid "class-blurring" effects consequential to the window-based computation of relevant measures. The procedure is validated based on results for the town of Bam, Iran, and compared with ground-based survey maps

In this paper we propose a novel procedure for urban road network extraction in high resolution SAR images. It is based on a multi-scale detection step including fusion of multiple features aimed at considering spatial high resolution as well as the spectral characteristics of the SAR images. Advantages over existing and previous extraction procedures are proved by comparison using data from different sensors and different test areas. I. INTRODUCTION High resolution SAR by Low Earth Orbit (LEO) satellites is going to have a deep impact on remote sensing data availability both because of the very short time between acquisitions and the spatial resolution, fine enough to monitor artificial structures. In turn, this will require more precise and efficient algorithms for the interpretation of this kind of SAR data. In fact, at high and very high resolution, natural and artificial objects must be individuated exploiting both their geometrical and spectral features, which show peculiar behaviors in SAR data. Following this idea, in this work we develop a decision fu- sion approach based on different detectors specifically tailored for road extraction from high resolution SAR data of urban areas. The approach exploits geometric a priori knowledge as well as spectral information about road materials. Street and roads in coarse SAR images may appear as dark or bright features, depending on their orientation. This is less true for high resolution SAR images, where roads are more-than-one-pixel wide: they most likely appear as dark, elongated areas, possibly with very bright sides. As a result, we may approach their detection and extraction by using geometrical analysis (1), looking for long edges, or by exploiting simpler multiple thresholding approaches (2), searching for dark, homogeneous areas. The proposed algorithm integrates both these approaches into a multi-scale feature fusion framework. Road candidate extraction in high resolution imagery usu- ally starts with road area detection, which of course may be obtained in optical images by looking for the spectral response of road materials. However, road class recognition in high resolution SAR data would imply complex segmentation algorithms based on data statistics. In this paper it is preferred to prove that multiple detectors may be enough to obtain good results. Therefore, in this paper a new extraction method is proposed, based on multiple feature detection and fusion , designed to be as automatic as possible and aimed at optimal junction preservation. The algorithm exploits spatial

Hyperspectral imaging offers the possibility of characterizing materials and objects in the air, land and water on the basis of the unique reflectance patterns that result from the interaction of solar energy with the molecular structure of the material. In this paper, we provide a seminal view on recent advances in techniques for hyperspectral data processing. Our main focus is on the development of approaches able to naturally integrate the spatial and spectral information available from the data. Special attention is paid to techniques that circumvent the curse of dimensionality introduced by high-dimensional data spaces. Experimental results, focused in this work on a specific case-study of urban data analysis, demonstrate the success of the considered techniques. This paper represents a first step towards the development of a quantitative and comparative assessment of advances in hyperspectral data processing techniques.

In this work, an all-digital synchronization recovery circuit is proposed for the uplink scheme of the ETSI BRAN Hiperaccess receiver. In particular a fast, open-loop algorithm, which exploits the TDMA burst preamble made of 16 or 32 cazac symbols, is tested by means of a link-level simulator operating at twice the symbol rate. At the receiver we suppose to use a fixed local oscillator and the symbol timing recovery is performed by means of digital interpolation. The fractional delay parameter for a fourth and sixth order Farrow interpolator is computed by means of a second order polynomial representation of the maximum likelihood function. After the recovery of the symbol timing, the phase offset is adjusted by simply taking the argument of the log-likelihood function. Performances of the proposed architecture are evaluated by measuring the symbol error rate and the synchronization error variances on AWGN channel.

In this letter, the problem of detecting urban road networks from high-resolution optical/synthetic aperture radar (SAR) images is addressed. To this end, this letter exploits a priori knowledge about road direction distribution in urban areas. In particular, this letter presents an adaptive filtering procedure able to capture the predominant directions of these roads and enhance the extraction results. After road element extraction, to both discard redundant segments and avoid gaps, a special perceptual grouping algorithm is devised, exploiting colinearity as well as proximity concepts. Finally, the road network topology is considered, checking for road intersections and regularizing the overall patterns using these focal points. The proposed procedure was tested on a pair of very high resolution images, one from an optical sensor and one from a SAR sensor. The experiments show an increase in both the completeness and the quality indexes for the extracted road network

This letter aims at the extraction of roads and road networks from high-resolution synthetic aperture radar data. Classical methods based on line detection do not use all the information available; indeed, in high-resolution data, roads are large enough to be considered as regions and can be characterized also by their statistics. This property can be used in a classification scheme. Therefore, this letter presents a road extraction method which is based on the fusion of classification (statistical information) and line detection (structural information). This fusion is done at the feature level, which helps to improve both the level of likelihood and the number of the extracted roads. The proposed approach is tested with two classification methods and one line extractor. Results on two different datasets are discussed.

In this paper, the problem of change detection from synthetic aperture radar (SAR) images is addressed. Feature-level change-detection algorithms are still in their preliminary design stage. Indeed, while pixel-based approaches are already implemented into existing, commercial software, this is not the case for feature comparison approaches. Here, the authors propose a joint use of both approaches. The approach is based on the extraction and comparison of linear features from multiple SAR images, to confirm pixel-based changes. Though simple, the methodology proves to be effective, irrespectively of misregistration errors due to reprojection problems or difference in the sensor's viewing geometry, which are common in multitemporal SAR images. The procedure is validated through synthetic examples, but also two real change-detection situations, using airborne and satellite SAR data over the area of the Getty Museum, Los Angeles, as well as over an area around the city of Bam, Iran, stricken in 2003 by a serious earthquake

A general processing framework for urban road network extraction in high-resolution synthetic aperture radar images is proposed. It is based on novel multiscale detection of street candidates, followed by optimization using a Markov random field description of the road network. The latter step, in the path of recent technical literature, is enriched by the inclusion of a priori knowledge about road junctions and the automatic choice of most of the involved parameters. Advantages over existing and previous extraction and optimization procedures are proved by comparison using data from different sensors and locations

In this paper, the problem of change detection in road networks from SAR images is addressed. Featurelevel change detection algorithms are still in their preliminary design stage. Indeed, while pixel-based approaches are already implemented into existing, commercial software, this is not the case for feature comparison approaches. As far as the change detection task is concerned, the availability of Synthetic Aperture Radar (SAR) data promises high potentialities, thanks both to the insensitivity of SAR imagery to atmospheric conditions and cloud cover issues and to the short revisit time planned for future SAR-based missions. Hence, multitemporal SAR imagery is expected to play a relevant role, for instance, with respect to ecological and environmental monitoring applications or to disaster assessment and prevention.

SAR images from space-borne platforms have proved to be helpful data for identification of oil spills and other surface anomalies, such as low wind areas, man-made targets, and natural films. The use of fractal dimension, which is related to the concept of surface "roughness", as a feature for classification, improves the detection of anomalies, since enhances texture discrimination. In the particular case of oil slicks, the surface tension of seawater is increased and the surface wave motion is significantly depressed. This effect relatively reduces the sea surface roughness, decreases the radar backscattered energy and enables oil slicks to be discernible from the radar image. Several algorithms may be applied for local fractal dimension estimation, but most solutions are tailored for specific applications and are characterized by estimation accuracies depending on the adopted image model and also on the value being estimated. This paper describes a decision-based fusion approach for local fractal dimension estimation of SAR images of the sea surface. Three different estimation algorithms are considered and the three resulting fractal maps are fused by means of a weighted average. The weights are calculated from the performance characteristics of the three algorithms measured on synthetic fractal surfaces. The experimental results carried out on ERS-2 SAR images prove the effectiveness of the proposed decision-based fusion approach

This paper is devoted to the description of a data set on a urban test site. This set is freely available to any researcher willing to test his/her algorithms on an urban area. The relevance of this set is tightly connected to the growing need of the urban remote sensing research community to compare algorithms and techniques.

In this paper, we analyze a multiple sensor data set corresponding to three-dimensional data coming from interferometric radar (InSAR) or laser ranging (LIDAR) measurements. LIDAR and InSAR are now mature technologies, and there are examples of their usefulness for urban area characterization. Unfortunately InSAR measurements show a serious disadvantage in describing built areas, due to problems derived from radar ranging. As a matter of fact, the possibility to have in the same area LIDAR data can reliably help in correcting all these effects. The advantage of LIDAR and InSAR joint use resides in exploiting the higher resolution offered by laser data and comes from the fact that LIDAR data is more expensive, and usually at the same cost we may obtain InSAR data on a much wider area than the one obtainable with a laser scanning survey

In this paper we investigate the use of SAR and multispectral sensors on board of the ENVISAT-1 satellite for urban remote sensing applications. We are interested mainly on the mapping capabilities of these two sensors and provide results for urban land use extraction using ASAR data and urban area definition using MERIS bands

Very high resolution hyperspectral data should be very useful to provide detailed maps of urban land cover. In order to provide such maps, both accurate and precise classification tools need, however, to be developed. In this letter, new methods for classification of hyperspectral remote sensing data are investigated, with the primary focus on multiple classifications and spatial analysis to improve mapping accuracy in urban areas. In particular, we compare spatial reclassification and mathematical morphology approaches. We show results for classification of DAIS data over the town of Pavia, in northern Italy. Classification maps of two test areas are given, and the overall and individual class accuracies are analyzed with respect to the parameters of the proposed classification procedures.

Provides a first assessment of a coregistration technique suitable for multiangle fine spatial resolution synthetic aperture radar (SAR) images. The technique is based on crossroad and road junction extraction and matching and exploits recently introduced road extraction routines for SAR data. These features are matched using relational and geometrical analysis. Results are encouraging and show the possibility to exploit multiangle SAR available from future airborne and satellite missions.

The aim of this paper is to investigate whether fractal maps extracted from sea SAR images are useful for discriminating the sea from other entities or anomalies. Fractal mapping consists of locally estimating the fractal dimension of the image. To this purpose four different methods based on covering and spectral analysis are proposed and compared when applied to real ERS1-2 GEC images. Wind falls, sea and line coast are well distinguishable in the fractal maps. This result clearly shows that the use of image fractal processing is a promising and powerful technique for identifying sea surface anomalies.

This paper is devoted to urban hyperspectral remote sensing. Very high resolution hyperspectral data are used to provide detailed maps of urban land cover, exploiting different classification tools. In particular, multiple classifications and spatial refinement step are used to improve the mapping accuracy. We show results on DAIS data over the town of Pavia, Northern Italy. The four flight lines over the area, kindly provided by DLR in the framework of the HySens project, are partially overlapping. This helps, besides the test of the classification procedure here presented, even to understand the advantages of combining different views of the same area.

In this work we present an image magnification technique aimed to improve the overall accuracy of hyperspectral data classification in an urban area. Furthermore, we discuss how techniques originally introduced for image enhancement in printers may be useful also for remote sensing applications. Finally, we compare different classifiers to look for the one able to exploit as much as possible the enhanced imagery. We find that fuzzy ARTMAP allows obtaining the best results.

In this paper we present and compare different techniques for the fusion of multitemporal SAR and multiband optical images. We consider both neuro-fuzzy and statistical approaches for the exploitation of the contextual information and the classification, and different schemes for the multisensor fusion. The proposed techniques are applied to a set of two multitemporal SAR and a Landsat multiband image of an urban area. Results show that it is possible to fully exploit the potentialities of the two sensors, by appropriately fusing their information. In particular, the proposed schemes are useful to retain at the same time the change detection capability and the best possible classification accuracy, thus they are of practical interest for civil protection applications.

We investigate the use of co-occurrence texture measures to provide information on different building densities inside a town structure. We try to improve the pixel-by-pixel classification of an urban area by considering texture measures as a means for block analysis and classification. We find some interesting hints concerning the optimal dimension of the window to be considered for texture measures, as well as the most useful measures. Moreover, we show that it is possible to use medium-resolution readily available satellite synthetic aperture radar images for a more refined urban analysis than previously shown.

In this paper, we present how we used neural networks (NNs) and a pyramidal approach to model the data obtained by a weather radar and to short-range forecast the rainfall behavior. Very short-range forecasting useful, for instance, for estimating the path attenuation in terrestrial point-to-point communications. Radial basis function NNs are used both to approximate the rain field and to forecast the parameters of this approximation in order to anticipate the movements and changes in geometric characteristics of significant meteorological structures. The procedure is validated by applying it to actual weather radar data and comparing the outcome with a linear forecasting method, the steady-state method, and the persistence method. The same approach is probably useful also for predicting the behavior of other meteorological phenomena like clusters of clouds observed from satellites.

We present some improvements to urban area characterization by means of synthetic aperture radar (SAR) images using multitemporal and multiangle datasets. The first aim of this research is to show that a temporal sequence of satellite SAR data may improve the classification accuracy and the discriminability of land cover classes in an urban area. Similarly, a second point worth discussing is to what extent multiangle SAR data allows extracting complementary urban features, exploiting different acquisition geometries. To these aims, in this paper, we show results on the same urban test site (Pavia, northern Italy), referring to a sequence of European Remote Sensing Satellite 1/2 (ERS-1/2) C-band images and to a set of simulated X-band data with a finer spatial resolution and different viewing angles. In particular, the multitemporal data is analyzed by means of a novel procedure based on a neuro-fuzzy classifier whose input is a subset of the ERS sequence chosen using the histogram distance index. Instead, the multiangle dataset is used to provide a better characterization of the road network in the area, overcoming effects due to the orientation of the SAR sensor.

In this paper, we derive two techniques for the classification of multifrequency/multitemporal polarimetric SAR images, based respectively on a statistical and on a neural approach. Both techniques are especially designed to exploit the spatial structure of the observed scene, thus allowing more stable classification results. Such techniques are useful when looking at medium- to large-scale features, like the boundaries between urban and nonurban areas. They are applied to a set of SIR-C images of a urban area, to test their effectiveness in the identification of the different classes that compose the observed scene. A lower and an upper bound to the classification performance are introduced to characterize their limits. They correspond respectively to pixel-by-pixel classification and to the joint classification of the pixels belonging to the different classes identified in the ground truth. The results achieved with the two approaches are quantitatively analyzed by comparing them to the ground truth. Moreover, a hybrid approach is presented, where the homogeneous regions identified through statistical segmentation are classified using a neurofuzzy technique. Finally, a quantitative analysis of the results achieved with all the proposed techniques is carried out, showing that their classification performance is much higher than the lower bound and reasonably close to the upper bound. This is a consequence of their effectiveness in the exploitation of the spatial information.

Fractal dimension of the sea surface is strictly related to its roughness, and may thus be helpful in determining the sea state, or where motion-damping oil spills are located. A possible way to determine the fractal dimension of the sea surface is that of performing a fractal analysis of remote sensing images, in particular satellite images, which have the advantage of observing a large area at one time. This paper aims to show the utility of fractal analysis of ERS-1 SAR images, and presents the results obtained by three different algorithms. The considered data was sensed by ERS-1 in the Mediterranean Sea at times and locations suitable for comparison with data coming from the Italian "Sistema Ondametrico Nazionale," an environmental measurement system including a number of buoys carrying accelerometers and communications instruments. The experimental results show some accordance between the buoy data and the ERS-1 fractal analysis outcome, but more data are required to provide statistical support to the conclusions.

This paper is devoted to the introduction of a fuzzy ARTMAP classifier based on a two-step approach. First, a pixel-by-pixel classification is performed, then a kernel-based refinement is applied to the output of the first step. The approach is applied to a multitemporal data set of a urban area, both to improve any single classification and to exploit the extra information carried by more images of the same area in different dates. We show that the fuzzy ARTMAP neural classifier is well suited to handle this kind of data. We discuss also the advantages deriving from the use of a multitemporal data set in urban areas, where many features are stable and so more data make their detection and recognition easier.

This paper deals with a feature fusion technique, especially implemented to characterize street networks extracted from multiple SAR images. The fusion approach may be useful in a number of ways: when applying different road extractors to the same data set, when using different pre-processing algorithms before applying the same road extractor to the same data set, or finally when applying the same road extractor to more images of the same area. The proposed approach applies "AND" and "OR" rules to the street nets to be compared. These operators are either hard or fuzzy ones, according to the way these networks were obtained. Therefore, the method is able to maintain the fuzzy reliability indicators obtained by fuzzy street extraction algorithms, but also to deal with differently classified segments and paths. We show through some examples how the proposed technique improves the results with respect to the "before fusion" street networks, and we also discuss future developments.

The problem of analysing and forecasting the motion of rain structures sensed by weather radar is mostly faced using approaches based on correlation of rain intensity values. Some approaches, however, consider rain structures as a base for the analysis. Of such approaches we considered a neural RBF-based one, of which we recently presented an improved version. The method we develop shows advantages over a linear prediction, while on the other side it is heavy and thus requires some tuning of the parameters to avoid exceedingly long processing times. In this paper we present some facts we discovered about time-saving compromises in tuning parameters and provide some rule-of-thumb guidelines for selecting their values.

In this paper we derive two techniques for the classification of multipolarimetric/multifrequency SAR images, based respectively on a statistical and on a neural approach. Both techniques are especially designed to exploit of the spatial structure of the observed scene, thus identifying homogeneous regions that can be jointly classified. Such techniques are useful when looking at medium to large scale features, like the boundaries between urban and non-urban areas. They are applied to a set of multipolarimetric/multifrequency SIRC images of a urban area, to test their effectiveness in the identification of built up areas. A quantitative comparison of the results achievable with the two techniques is carried out, showing a similar behavior, even if the statistical approach tends to achieve better performance.

The paper relates to some developments in modeling attenuation time series, measured on Ka band satellite links during rain events, by means of sample generators using chaotic dynamical systems. Weighted sums of Lorenz strange attractors are first used to model target sequences by themselves ("global approach"), improving procedures already introduced. Then, similar generators are used to model only low-pass versions of the same sequences, while hidden Markov models (HMM) or shaped random number series are used to mimic the complementary high-pass sequences.

Previous works have introduced models based on deterministic chaos equations, aiming to simulate transmission processes over mobile radio digital channels at the level of the post detection error stochastic process and to reproduce their renewal or nonrenewal behaviors. In particular bursts and clusters of errors are generated by these models, exploiting the correlation between successive points sampled from the trajectories of suitable strange attractors. In this paper chaos equation models are reviewed and relevant results are discussed. Different approaches have been tested, using one or more attractors, and calling for light or heavy preprocess procedures to cope with the statistical characteristics of target gap-time series. Models derived from hidden Markov chains have been implemented to provide comparison of results. Target sequences were supplied by simulation of digital enhanced cordless telecommunications (DECT) channels or derived from reception of DECT signals by a mobile unit in indoor environments.

In this paper, we describe a method to obtain a reliable set of elevation data suitable for data validation on the Shuttle Radar Topography Mission (SRTM), starting from laser scanning measurements on an urban test site: Pavia, Northern Italy. The elevation dataset is obtained through extraction of digital terrain models. The source digital surface model is first filtered by means of a lowpass or morphological kernel. Then, buildings are suppressed through analysis of the height histogram. Finally, a lowpass filter suppresses the surviving elevation artifacts. We show that, starting from a digital surface model at 1-m ground resolution, we end up with a digital terrain model that can be used as a ground truth for SRTM topographic analysis of an urban area.

The paper relates on experiments in deriving models for satellite link channels in the Ka band using nonlinear deterministic chaos equations. The trajectories of Lorenz attractors are sampled at suitable time rates and combined to obtain attenuation samples of the fading channel, following procedures suggested by previous experience in modeling the behavior of time error series in digital communication channels. Optimization cost function structures are discussed.

The authors work on meteorological satellite image archives and provide a novel and useful query-by-shape tool. To this aim, they first present the point diffusion technique (PDT), a fast and efficient method for shape similarity evaluation. Thanks to its very structure, this approach is suitable to handle objects whose shape is not well defined and can be represented by a set of sparse points. PDT is thus suitable for application to similarity-based retrieval from remotely sensed image archives, where shapes are hardly defined but are still among the major features of interest. Moreover, they prove here that PDT is almost as effective as more standard procedures for shape-based database queries, although significantly faster. In other words, it manages to combine retrieval speed and precision, the features of greatest importance for a first remote sensing data prescreening in many applications. Archives of meteorological satellite images are typical examples of very large-sized, remote sensing-based databases with a special attention for shape features. Each meteorological satellite produces terabytes of data every day, a large part of which is not immediately analyzed and ends being stored in archives. The application of PDT to such a database is presented and discussed, and a comparison with a standard method developed for meteorological shape analysis is provided.

The authors present a fuzzy approach to the analysis of airborne synthetic aperture radar (SAR) images of urban environments. In particular, they want to show how to implement structure extraction algorithms based on fuzzy clustering unsupervised approaches. To this aim, the idea is to segment first the sensed data and recognize very basic urban classes (vegetation, roads, and built areas). Then, from these classes, we extract structures and infrastructures of interest. The initial clustering step is obtained by means of fuzzy logic concepts and the successive analyses are able to exploit the corresponding fuzzy partition. As a possible complete procedure for urban SAR images, they focus on the street tracking and extraction problem. Three road extraction algorithms available in literature (namely, the connectivity weighted Hough transform (CWHT), the rotation Hough transform, and the shortest path extraction) have been modified to be consistent with the previously computed fuzzy clustering results. Their different capabilities are applied for the characterization of streets with different width and shape. The whole approach is validated by the analysis of AIRSAR images of Los Angeles, CA.

The authors propose a pattern recognition approach to analyze subsurface radar images and recognize the hyperbolic signatures produced by targets. After enhancing these signatures using a wavelet denoising procedure, pixels are grouped into significant hyperbolic shapes by fuzzy clustering. The approach also provides a validation measure for each recognized shape using so-called "shell thickness". Excellent experimental results justify the use of this algorithm for automatic interpretation of subsurface radar images.

The task of extracting significant built structure in digital surface models (DSM) is analyzed. The original data are obtained by means of interferometric SAR or LIDAR techniques and have different resolution and noise characteristics. This work aims to make a comparison of what (and how precisely) it is possible to detect and extract starting from these models, taking into account their differences but applying to them the same planar approximation approach. To this aim, data over Los Angeles and Denver is considered and evaluated. The results show that LIDAR data provide a better shape characterization of each building, and not simply because of their higher resolution. Indeed, less accurate results obtained starting from radar data are mainly due to shadowing/layover effects, which can be only partially corrected by means of the segmentation procedures. However, better results than those already presented in the literature could be achieved by using the IFSAR data correlation map.

The authors present a complete procedure for the extraction and characterization of building structures starting from the three-dimensional (3D) terrain elevation data provided by interferometric SAR measurements. Each building is detected and isolated from the surroundings by means of a suitably modified machine vision approach, originally developed for range image segmentation. The procedure is based on a local approximation of the 3D data by means of best-fitting planes. In this way, a building footprint, height and position, as well as its description with a simple 3D model, are recovered by a self-consistent partitioning of the topographic surface reconstructed from interferometric radar data.

A processing chain for the spatial analysis of the data recorded by a ground penetrating radar (GPR) is presented. In particular, the detection and localization of pipes is implemented by exploiting the a priori knowledge that a buried cylinder gives rise to a hyperbolic signature in GPR images. The image interpretation is performed by a suitably trained simple neural detector after some preprocessing steps aiming toward the enhancement of the buried objects' signatures. The algorithm has been tested on actual GPR images and compared with the information extracted by a trained human operator, and the agreement is extremely satisfying. Moreover, the possibilities and advantages to exploiting some sort of "spatial diversity" by combining the analysis of data simultaneously recorded by different antennas are presented and discussed.

The neural network approach is applied to the detection of cylindric objects as well as their geometric and electrical characteristics inside a given investigation domain. The electric field values scattered by the object and available at a small number of locations are fed into the network, whose output is the dielectric permittivity, and the location and radius of the cylinder. The results are evaluated using different sets of testing data, and the dependence of the various output parameters to the input are considered. The algorithm performance shows that the approach is able to solve the inverse scattering problem quickly. This may be useful for real-time remote-sensing applications.

Images from meteorological satellites or weather radars must be often interpreted at a higher level than simple pixel classification, and shape analysis with reliable and fast methods may be necessary. For instance, it is useful to analyze the temporal evolution of a rain event by means of a reliable tracking of the rain patterns at different scales. However, this task of tracking objects continuously changing their shapes is challenging. In this paper, the author shows how the application of a recently introduced numerical technique, called boundary integral-resonant mode expansion (BI-RME), to weather radar and meteorological satellite data could be used to achieve more information about the evolution in time of rain patterns or other meteorological structures of interest. In order to demonstrate the efficiency and robustness of the approach, several examples of image processing are considered. Applications to both operational tracking of clouds to produce wind fields and hurricane tracking are presented.

Source modelling is one of the fundamental issues in studying the performance of networks under different traffic conditions. Previously, emphasis has been put on modelling the output of digital video sources to determine the actual load and its temporal correlation. In this paper we propose a new modelling approach based on the use of chaotic equations to generate simulated MPEG sequences. The performances of an ATM node carrying video traffic are observed using both "real" and "chaotic" input video sources. We show that the proposed model is accurate enough to capture the loss-delay characteristics of the node and can thus be used as an alternative to traditional probabilistic models.

The so called point diffusion technique is introduced. It allows one to quickly evaluate the similarity between two figures defined by "clouds" of points. The new technique is compared with widespread modal matching approaches, showing similar results with respect to the efficiency of the queries in an image database, and using sensibly less CPU-time.

We present the application of a simplified shape analysis technique based on a modal representation of the object shape, and which is useful for improving the efficiency and effectiveness of shape-driven searches in image databases. The proposed method computes the representation of an object by means of modes very similar to the deformation modes of a mechanical system, but in a numerically more stable way than the usual finite-element method approach. Moreover, to make the technique for the visual search more effective, many different definitions of similarity indexes are introduced and discussed. The problems related to the comparison between objects represented by a very different number of feature points are also discussed. Finally, to prove the effectiveness of the approach, the indexes are studied in a simple case study (a small database of character shapes). However, their performance on a larger image database is also addressed, as well as the ability of the method to efficiently assess the problem of retrieving images similar to a user-defined sketch.

A system devoted to ship traffic control in a harbour environment is proposed, where optic flow approach and monocular image sequences are used. In each frame the scene is segmented, moving and still objects are found, and the movement of each ship is completely tracked. Partial and/or total occlusions are correctly handled by means of suitable grouping algorithms. Quantitative motion estimation is obtained by an iterative procedure that extract precise 3D information from the monocular sequence. The implemented version of the system is able to monitor and control a harbour environment with a substantially low computational effort.

We present a fast algorithm to extract a shadow model from a monocular colour scene exploiting the hue, luminosity, saturation (HLS) colour components. The method allows one to recover target shapes in diurnal scene for improved identification and it is based on the definition of a global bitmap model and a more particular strip bitmap model to identify shadow regions. Each pixel in the image is then classified as shadow or target by a comparison with these models.

This paper proposes a new algorithm which applies perceptual grouping to track discontinuous chains of symbols in digitized maps. The procedure is based on an artificial intelligence kernel that supervises three different auxiliary processes: the search strategy generation module, responsible for the strategy to scan pixels; the symbol detection module that extracts the recognized symbols; the cost function evaluation module that assigns a global quality index to each symbol by considering the whole course of the line. Selected Gestalt rules are used to optimize the grouping procedures.

An efficient vanishing point detection algorithm for non-structured scenes is proposed. The technique relies on a voting scheme based on well-understood rules for the grouping of lines. Infinite distance vanishing points are previously detected by a basic algorithm to simplify the successive search.

When processing multidimensional remote sensing data, one of the main problem is the choice for the appropriate number of clusters; despite of the great number of good algorithms for clustering, each of them works properly only when the appropriate number of clusters is selected. As adaptive versions of the K-means, competitive learning (CL) algorithms also have a similar crucial problem; various efforts to improve the performance of CL were made with the introduction of frequency sensitive competitive learning (FSCL) and rival penalised competitive learning (RPCL). We present an improvement of the RPCL algorithm well adapted to work with every kind of real clustering data problems. The basic idea of this new algorithm is to introduce a competition also between the weights. The algorithm was tested on multiband images with different weights initial position, giving similar results.