Hurricanes cause significant damage to communities, necessitating rapid and accurate damage assessment to support timely disaster response. However, image-based deep learning models for hurricane-induced damage assessment face substantial challenges due to domain shifts across different hurricane events, and the restricted availability of labeled data for each disaster further complicates this task. In this study, we propose a novel domain-adaptive deep learning framework that mitigates the domain gap while requiring minimal labeled samples from the target domain. Our approach integrates a self-supervised learning (SSL) pretext task to enhance feature robustness and leverages a novel Bilateral Local Moran’s I module to improve spatial feature aggregation for damage localization. We evaluate our method using aerial datasets from Hurricanes Harvey, Matthew, and Michael. Experimental results demonstrate that our model achieves more than 5% improvement in damage classification accuracy over baseline methods. These findings highlight the potential of our approach for scalable and efficient hurricane damage assessment in real-world disaster scenarios.

Texture feature extraction is very important in landuse land cover(LULC) classification of satellite images. Through this paper, a new method for texture feature extraction is presented which uses singular value decomposition(SVD) and gray level cooccurance matrix (GLCM). Here we test for capabilities of the singular values thus generated for classification of textures. We also try to find out if these singular values can be used as a substitute for Haralick texture features. In this proposed method sample images are multi-thresholded to reduce the dimensionality. GLCM is generated for each image-patch after applying the thresholds. Later, the SVD decomposition of GLCM provides singular values which are used as a feature vector for classification of the image patches. We evaluated our proposed technique based on three criteria a) images from totally different classes b) images from same class but with different textures c) images from same class, same texture but different orientation. We classified the images using minimum distance to mean(MDM), SVM using radial bias kernel (cSVM) and SVM using quantum kernel (qSVM). We used IBM gate-based qiskit to generate a quantum kernel and classify using qSVM. From the experiment we conclude that singluar values are good and stable as textures features and greatly enhance the texture classification. Singular Values along with quantum kernel have produced very promising results.

Critical infrastructure (CI) plays a pivotal role in supporting daily life, encompassing vital sectors such as Healthcare, Transportation, and Power systems. Failure of CI can significantly impede everyday activities, making it essential to understand and address the interdependencies among these infrastructures. Failures can arise from both natural and manmade events, often leading to cascading effects across CI sectors. To mitigate capital loss and adversity during such failures, the efficient utilization of limited resources becomes crucial. This paper introduces a real-time decision-making system that incorporates local and global factors impacting CI and enables effective resource allocation. In this study, we focus on the failure of a specific CI sector, healthcare, caused by a flood event. To simulate this scenario, we create an environment that captures interdependencies between CI sectors, while generating flood events through randomly distributed water levels over time. To optimize decision-making, we employ a Reinforcement Learning (RL) based agent trained using Deep Q learning. The trained agent suggests critical decisions that enhance the utilization of limited resources, thereby extending the system's survivability. To provide a comprehensive view of the system's state and actions recommended by the learned agent at each time step (t), we develop a user interface. This interface displays the environment's states and facilitates the visualization of alternative CI protection strategies during catastrophic events like floods. Such simulation environments empower decision-makers with vital capabilities to make informed choices regarding resource allocation in critical scenarios.

Quantum Machine Learning (QML) is a new interdisciplinary branch that combines quantum computing with machine learning. It is emerging as an alternative to classical machine learning which exploits the quantum mechanical properties of entanglement and superposition to express the hidden patterns in the data. This reduces computational resources as well as the time required for processing. In this study, we tried to address the challenge of landuse land cover classification for classes with similar spectral signatures and small texture dissimilarity. Totally 6 textural features are extracted from a multispectral image using grey level co-occurrence matrix applied on 1st PC component to classify images into the sub-classes(residential, highway, and industrial) of the main built-up class. We used IBM gate-based qiskit to generate a quantum kernel and classify the images using QSVC. Quantum kernels are very expressive when compared to their classical counterparts and can learn complex data more efficiently. The overall accuracy of classification by QSVC is comparable to that of the classical SVC. We summarize our results by saying that QSVC performs better than SVC.

Light Detection and Ranging (LiDAR) is an active remote sensing technique that uses pulsed lasers to sense the surrounding environment. It works on the principle of Time of Travel (ToT) for acquiring highly accurate and high spatial resolution 3D information about the surrounding environment. LiDAR data possesses immense potential for (near) real-time applications (e.g., forestry, disaster, border security, etc.) Real-time applications demand quick analysis of the data and cannot wait for the entire data pertaining to the area of interest to be captured before producing useful insights. Thus, 3D LiDAR points should be processed as and when they are captured in the form of a continuous stream. Due to the lack of prior knowledge about the (near) real-time data and the underlying distribution, in this work, unsupervised stream mining approaches have been adapted for the analysis of streaming geospatial 3D LiDAR data. By applying different unsupervised data mining (clustering) algorithms on a huge set of 3D LiDAR data points, we have evaluated the quality of clustering based on different evaluation metrics.

In a traditional Machine Learning setting, it is assumed that the training and test dataset belong to the same feature space and the same distribution. For practical real-world scenarios, this is not always true. The model which is trained on the source dataset might give poor results on the target dataset (during testing/inferencing). We address this challenge using Unsupervised Domain Adaptation which deals with the situation in which the network is trained on the labeled source domain data and unlabeled data belonging to a different but related target domain and it gives improved results when tested on the target domain data. Also, generating data considering two different but related domains into account is a challenge. In this work, we address this challenge by proposing a framework for developing a cross-domain Generative Adversarial Network (GAN) with the aim of achieving unsupervised domain adaptation which applies to LiDAR data appertaining to extreme events. The dataset used for the study is the pre- and post-earthquake LiDAR (Light Detection and Ranging) dataset belonging to the Kumamoto earthquake of 2016. A comprehensive comparative analysis is conducted in this study to examine the efficacy of urban damage classification, both with and without the integration of unsupervised domain adaptation. We envisage that our framework can be used for performing urban damage classification in an efficient manner with increased accuracy.

There has been an increase in the adoption of Linked Data and subsequently representing data in the form of knowledge graphs across a wide spectrum of domains. There has also been significant interest in the remote sensing community to publish Earth Observation data in the form of Linked Data. As the geospatial Linked Data cloud on the internet grows, there arises a need for efficient methods of exploratory analysis of such information-rich geospatial knowledge graphs. Knowledge graph representation of remote sensing scenes has proved to add significant value for effective mining of implicit information in addition to seamless integration with other data sources. This work is geared towards visual exploration of semantically enriched Remote Sensing Scene Knowledge Graphs (RSS-KGs). In this paper, we propose and implement an interactive web-based interface to visually explore and interact with RSS-KGs using Cesium. The proposed interface seeks to visualize the know ledge graph in the form of nodes and edges, mapped over the remote sensing scene consisting of different land use land cover regions and their inferred characteristics in addition to their spatial relationships with one another. It is envisaged that visualization in the form of nodes and edges would aid in visually validating the spatial relations in the knowledge graph, thus enhancing the understanding of the geospatial knowledge graph from the end user perspective. We demonstrate the efficacy of the interface through the visual exploration of an enriched geospatial knowledge graph of a remote sensing scene captured during an urban flood event.

Remote sensing lidar point cloud dataset embeds inherent 3D topological, topographical and complex geometrical information which possess immense potential in applications involving machine-understandable 3D perception. The lidar point clouds are unstructured, unlike images, and hence are challenging to process. In our work, we are exploring the possibility of deep learning-based question-answering on the lidar 3D point clouds. We are proposing a deep CNN-RNN parallel architecture to learn lidar point cloud features and word embedding from the questions and fuse them to form a feature mapping for generating answers. We have restricted our experiments for the urban domain and present preliminary results of binary question-answering (yes/no) using the urban lidar point clouds based on the perplexity, edit distance, evaluation loss, and sequence accuracy as the performance metrics. Our proposed hypothesis of lidar question-answering is the first attempt, to the best of our knowledge, and we envisage that our novel work could be a foundation in using lidar point clouds for enhanced 3D perception in an urban environment. We envisage that our proposed lidar question-answering could be extended for machine comprehension-based applications such as rendering lidar scene descriptions and content-based 3D scene retrieval.

Inability to respond to the growing trend of COVID -19 cases and the study and analysis of Healthcare Critical Infrastructure interdependencies during COVID-19 pandemic scenario is relatively new. One of the most frequently identified shortfalls in knowledge related to enhancing Healthcare Critical Infrastructure (HCI) preparedness during the COVID-19 pandemic scenario is the inability to forecast the growth trend of COVID-19 cases in a geographic area and incomplete understanding of interdependencies between Critical infrastructures related to HCI. As the number of cases surges at a healthcare facility, the facility, and its interdependent CI services should be prepared to handle the susceptible stress. The goal of the paper is to be able to predict the growth trend of COVID-19 cases using Spatiotemporal Long Short-Term Memory (ST-LSTM) for a geographic area. Based on the predicted growth trend of the COVID-19 cases a Multi-Agent Deep Reinforcement Learning (MADRL) simulation model will provide an accurate representation of healthcare critical infrastructure characteristics, operations, and interdependencies services. The Real-time information simulation would help frontline workers, government agencies, and disaster and emergency response personnel to respond to the question, ‘what if something else happens during the COVID-19 Pandemic?

In the early hours following the earthquake, supporting humanitarian actions like rescue operations and relief distribution is the primary objective of the rescue managers. The damage mapping can be performed using reliable data that can be obtained from high-resolution satellite imagery but obtaining satellite imagery can be challenging for some days post disaster due to revisit time. Considering the disaster response timing, Unmanned Aerial Vehicles (UAV) are used because ground transportation systems are ineffective due to road blockage. In this work, we make use of Light Detection and Ranging (LiDAR) 3D point cloud data obtained for Haiti Earthquake. The focus of our work is to develop and implement an approach for LiDAR data classification to enable Earthquake damage mapping and detection. This is obtained by running our deep learning network on NVIDIA Jetson Nano embedded supercomputing platform. This approach takes the advantage of embedded High-Performance computing and low power consumption capabilities of Jetson Nano which enhances the classification and promotes rapid response which is the key to manage post-disaster activities. Jetson Nano is a feasible option which provides a GPU architecture that is optimized for running energy-aware deep learning models and which generates the results in real or near-real time. We envisage that our work could be extended to perform near real-time classification of LiDAR point clouds in a post earthquake scenario.

The advent of the Internet of Things (IoT) has led to the generation of tremendous amounts of data from various sources. Cloud based systems are effective in storage and application of machine learning algorithms on such datasets. However, in some cases it is important to enable real time processing for making immediate decisions. There are many applications which require instantaneous analysis of the generated data for remedies in event of an anomaly. Data associated with such use cases remains significant only for a short duration of time. Various electronic sensors, e.g. Temperature, Moisture, Air Quality, Pressure, Wind Velocity etc. present in a Wireless Sensor Network generate streams of values. It can be processed using pipelines which provide prompt and quick analysis for decision making. Stream Processing Systems can be helpful in such cases as they analyse data streams within a few milliseconds to a few seconds. In this paper, we discuss an event based processing of streaming data from air pollution sensors to create a real time anomaly detection system. To reduce the delays associated with the generation of alarms in our pipeline, Apache Foundation's Stream Processing Tools, Kafka and Flink were used for operations on our streams. To further accelerate the process, all the analysis is conducted on an embedded edge computing gateway device rather than sending data to the cloud for batch processing. The results are obtained in the form of a geographical map visualization using ELK stack. The map highlights the coordinates of the location with an unhealthy air quality index in real time.

With an increase in Geospatial Linked Open Data being adopted and published over the web, there is a need to develop intuitive interfaces and systems for seamless and efficient exploratory analysis of such rich heterogeneous multi-modal datasets. This work is geared towards improving the exploration process of Earth Observation (EO) Linked Data by developing a natural language interface to facilitate querying. Questions asked over Earth Observation Linked Data have an inherent spatio-temporal dimension and can be represented using GeoSPArql. This paper seeks to study and analyze the use of RNN-based neural machine translation with attention for transforming natural language questions into GeoSPArql queries. Specifically, it aims to assess the feasibility of a neural approach for identifying and mapping spatial predicates in natural language to GeoSPARQL's topology vocabulary extension including - Egenhofer and RCC8 relations. The queries can then be executed over a triple store to yield answers for the natural language questions. A dataset consisting of mappings from natural language questions to GeoSPArql queries over the Corine Land Cover(CLC) Linked Data has been created to train and validate the deep neural network. From our experiments, it is evident that neural machine translation with attention is a promising approach for the task of translating spatial predicates in natural language questions to GeoSPArql queries.

The paper proposes a Multi-Agent Deep Reinforcement Learning (MADRL) simulation model that is useful in understanding the status of Critical Infrastructures (CI) during extreme events. The simulation model can be used to understand the spatiotemporal nature of the event and evaluate and predict the propagation of cascading failure scenarios in the critical infrastructure network. Multi agent-based modeling is performed by interconnecting multiple agents, which are autonomous computational entities. Geospatial based intelligent agents are developed, such that each agent registers with a CI such as a Healthcare infrastructure agent, Transportation agent, etc. These agents check for an infrastructure state change (e.g. the roads which lead to the hospital are blocked due to debris), and if there is a state change then they would reason about the impacts of these events upon other dependent infrastructures. Deep reinforcement learning approach helps the geospatial based CI agents in making a rapid and an optimal decision based on its spatiotemporal environment, during a flood event. The utility of the approach is evaluated using a real-world case study. Real-time information simulation would help disaster response personnel to respond to the question, `what if something else happens?

Real-time embedded vision tasks require extraction of complex geometric and morphological features from the raw 3D point cloud acquired using range scanning systems like lidar, radar etc. and depth cameras. Such applications are found in autonomous navigation, surveying, 3D mapping and localization tasks such as automatic target recognition (ATR). Typically, a dataset acquired during surveying by remote sensing lidar scanners, known as point cloud, is (1) huge in size and requires a big chunk of memory for processing at a single instance and, (2) experiences missing information due to rapid change in orientation of the sensor while scanning. In our work, we are addressing both the issues combinedly by proposing an online point cloud super-resolution approach for translating a low dimensional point cloud to a high dimensional dense point cloud by learning dictionaries in the low-dimensional subspace. We are presenting our approach for an urban road scenario by reconstructing dense point clouds of 3D objects and comparing results based on PSNR and Hausdorff distance.

Very High Resolution (VHR) remote sensing optical imagery is a huge source of information that can be utilized for earthquake damage detection and assessment. Time critical task such as performing the damage assessment, providing immediate delivery of relief assistance require immediate response; however, processing voluminous VHR imagery using highly accurate, but computationally expensive deep learning algorithms demands the High Performance Computing (HPC) power. To maximize the accuracy, deep convolution neural network (CNN) model is designed especially for the earthquake damage detection using remote sensing data and implemented using high performance GPU without compromising with the execution time. Geoeye1 VHR disaster images of the Haiti earthquake occurred in year 2010 is used for analysis. Proposed model provides good accuracy for damage detection; also significant execution speed is observed on GPU K80 High Performance Computing (HPC) platform.

In disasters, since time is of the essence, quick decision making based on actionable insights is desired. In our earlier work, we have demonstrated that the spatial relationships-based queries can play a vital role in the disaster response phase. However, we found that the utilization of spatial relationships rules (i.e. encoded spatial knowledge) via rule reasoning process do not scale well with the increased number of image regions. Most of the available Resource Description Framework (RDF) triplestores do not support rule reasoning due to the computational complexity and undecidable nature of the rule reasoning process. In this paper, we propose an alternative approach for utilizing spatial knowledge encoded in the form of spatial relationship rules. The proposed approach reformulates the spatial query by expanding it with the configuration encoded in the corresponding spatial relationship rule. The preliminary results are promising and show the applicability of the proposed approach during the time critical events such as flood disaster.

Natural disasters such as floods, earthquakes, hurricanes, etc. have a huge impact on a society-causing destruction of life and property in their wake. During disasters such as flood, it is crucial to understand the dynamics of the situation as it occurs for effective response. In this paper, we address the problem of satellite image classification for urban floods using deep learning. We propose an encoder-decoder neural network based on the Efficient Residual Factorized Convnet(ERFNet), for multi-class segmentation of urban floods from multi-spectral satellite imagery. The ERFNet architecture capitalizes on skip connections and one dimensional convolutions to achieve the best possible trade-off between accuracy and efficiency. Since time is of essence during a disaster, the choice of the ERFNet architecture on a high performance computing (HPC) platform is apt. Satellite imagery from WorldView-2 of floods in Srinagar, India during September 2014 have been used for this study. The tool `markGT' has been developed to assist end-to-end annotation of satellite imagery. The urban flood dataset used for this study has been generated using markGT. The proposed deep learning model over urban flood satellite imagery gives promising results on Nvidia Tesla K80 GPU. We envisage that the proposed model could be extended and improved for real-time classification of urban floods, thereby aiding disaster response personnel in making informed decisions.

Disasters such as earthquakes, floods, landslides etc. create great economic and social loss by destroying the balance of life and property and create chaos. In the wake of a disaster, it becomes very significant to take real-time and on-the-fly actions to minimize the effects of the event. Remote Sensing data acquired through airborne or spaceborne platforms is usually huge in size and requires huge time in generating actionable insights during the disaster scenario. In this work, we propose a two-fold analysis of the Very High Resolution (VHR) satellite imagery based on Compressive Sensing (CS) and Deep Learning. We propose employing a deep learning approach for inferencing over compressed sensing satellite imagery. We hypothesize that this could be beneficial in generating real-time actionable insights during a catastrophe. In our work, we are using the satellite imagery from GeoEye-1 of Haiti Earthquake. Our objectives are: (1) To generate CS images for 75%, 50%, and, 25% sampling on the sparse space and (2) To develop a deep learning pixel-level classification model based on the UNet architecture using the original and reconstructed images. The UNet architecture has shown promising results for pixel-level classification in the recent literature. We envisage to combine both the objectives into an end-to-end learning framework for on-board processing which we foresee would be of great significance in various applications for rapid disaster management response.

Numerous remote sensing applications - flood monitoring, forest fires monitoring, earthquake analysis etc. require users to query satellite images based on their content. Such requirements have led to the evolution of Content-based Image Information Mining Systems over the last decade. Recent developments in the area of Image Information Mining(IIM) are geared towards bridging the gap between low level image features and higher-level semantics. This research focuses on improving the semantic understanding of a remote sensing scene during the flood disaster from a spatio-contextual standpoint. During a flood occurrence, it is crucial to understand the flood inundation and receding patterns in context to the spatial configurations of the land-use/land-cover in the flooded regions. This study focuses on bridging the spatio-contextual semantic gap in understanding of the remote sensing imagery during a flood, thereby attempting to improve the machine interpretability of a flood remote sensing imagery. In this regard, the Flood Scene Ontology (FSO) has been developed to mine the topological and directional knowledge in context to the flood disaster phenomenon. The FSO is envisaged to form the basis for developing applications that would utilize the spatio-contextual semantics of the flood disaster to aid in the Disaster Assessment and Management process. This paper describes the conceptual framework that was developed to address the same.

Jetson TK1 is the first mobile processor from NVIDIA having similar features and architecture as that of a modern desktop GPU and still using low power from a mobile chip. Therefore, Jetson TK1 runs the same CUDA code (running on desktop GPU) with similar level of performance. Also, with the dawn of GPU technology, it has become possible to perform tasks (that are computationally intensive) in realtime or near-real time. In the agricultural domain, retrieving the biophysical parameters of the crop is important as it provides insights into the plant growth status. Inversion of the Radiative Transfer Model enables to obtain these parameters. However, such a process is highly computationally intensive. The focus of this work is to develop and implement an approach that takes the advantage of embedded High-Performance Computing (HPC) capability of Jetson TK1 to significantly improve the inversion process of a Radiative Transfer Model. The experimental results show that Jetson TK1 based biophysical parameters estimation gives significant speedup, which opens-up the possibility of having a Jetson based embedded platform for on-board biophysical parameters estimation in the future. In such a scenario, where there are constraints related to energy and power, Jetson TK1 can become a practicable option by providing a GPU based architecture for running energy-aware computationally intensive algorithms in parallel for processing the data, and generating the results in real-time or near-real time while taking care of the power usage.

During an extreme event such as flood disaster, it is important to study flood inundation and receding patterns to understand the dynamic spatio-temporal behavior of flood. In addition to the general change detection techniques in RS, a proper conceptualization of `change' during flood disaster is necessary to model its dynamic behavior. This motivated the development of Ontology, which is able to capture the dynamically evolving phenomenon. This Ontology is envisaged as a precursor for developing applications that integrate the spatio-temporal dimensions of a dynamically evolving system such as floods. This paper describes the conceptual framework that was developed to address the same.

Floods can have a devastating effect on critical infrastructures, especially in the case of cascading effects among multiple infrastructures such as the Healthcare Critical Infrastructure (CI), Electricity CI, Water supply CI, etc. To ensure a resilient Healthcare Critical Infrastructure, understanding the vulnerabilities and analyzing the interdependencies on other critical infrastructures is important. In order to enable real time situational awareness for operational risk management one needs to be aware of the wide range of events that can unfold during a disaster and their effect on CIs. In this paper, we present a dynamic simulation model based on a Petri net modeling approach integrated with a geographic information system and Geo-visual analytics for detecting cascading failures and visualization of CI interdependencies during a flood disaster scenario to enable timely and efficient response.

Voluminous data (Multispectral, Hyperspectral) from Variety of sensors (Airborne sensors, space borne sensors) with Velocity (high temporal resolution) when used for decision making to support natural disasters such as earthquakes, floods, oil-spills etc., for near real time accurate responses, is a problem that needs Big Data Analytics. To gain rapid insight from this big data, high performance computing (HPC) with some scalable solution that reduces the execution time are in extreme demand. To serve this real time need, scalable hybrid parallelism approach based on state of art multi-core GPUs and Message Passing Interface (MPI) is explored for analyzing remote sensing disaster data. Spatio-temporal remote sensing data of oil-spill at Gulf of Mexico captured by LANDSAT 7 ETM+ is considered for analysis. The core objective includes performance evaluation of the analysis process across various parallel implementation platforms.

Interoperability is a key concern in geospatial domain. Approaches based on Geo-ontology which represents the semantics of the geospatial information are currently popular. However, there still remains the problem of reconciliation of different geo-ontological representations. Hence, many geo-ontology matching systems have been developed to harmonize and integrate various information sources for resolving the heterogeneity and achieve interoperability. However, deriving the degree of interoperability between two information sources is a challenge. Measuring the degree of interoperability helps in understanding, the readiness or the amenability to communicate across similar representations, thus identifying those information sources (out of many similar ones) that can be integrated. This paper describes a Geo-ontology matching framework to measure the degree of interoperability.

The recent development in sensor technology shows the unprecedented growth of Remote Sensing (RS) data archives-Big Data. However, this growth in RS archives has resulted in many processing challenges. The three V's of big data- Volume, Velocity and Variety is highly relevant in situations such as flood, earthquake disaster, where real/near real time processing of data from different RS data sources is vital to deploy rescue operations. In this work, we have demonstrated a high-performance analytics approach- Message Passing Interface (MPI) along with the emerging Graphics Processing Units (GPUs) (i.e. hybrid MPI+GPU) technology to overcome the big data processing limitation. The different processing/analysis stages of our Spatial Image Information Mining (SIIM) system are parallelized using the above approach. The experimental results for parallel segmentation process show the applicability of MPI+GPU hybrid approach in disaster scenario.

There is a need for rapid response during disasters. However, there is a paucity of training data which leads to classification models that do not generalize well. If the pre disaster data is used to augment the training data, the models perform poorly due to statistical distribution differences between pre and post disaster conditions. Also, it is challenging to analyze large areas for identifying the disaster affected regions by visual image interpretation techniques. Using the limited available ground truth, during post disasters, a domain adaptation (DA) approach is used to study the post-earthquake situations. Further, knowledge about the spatial relationships with adjacent regions is integrated with the DA approach to refine the classification. The results were compared to traditional classification methods and were found to achieve higher accuracies with smaller training sample sizes.

High resolution satellite imagery is useful for disaster management activities such as damage assessment, immediate delivery of relief assistance etc. The process of analyzing Satellite imagery involves extraction of optimal features that closely represent the damaged areas. Accuracy of the analysis depends on the efficiency and robustness of selected features. Scale invariant feature transform (SIFT) enables to extract features, which are scale and rotation invariant. It provides robust features even in cluttered and partially-occluded images (such as those images that are obtained from a post disaster scenario). SIFT is robust at the cost of multiple stages involved in making features scale and rotation invariant, which is a time intensive process to apply on high resolution imagery. In general, there is a need to synthesize large amount of high-resolution, high temporal satellite data for disaster management applications to enable near real time response. However, this task is computationally intensive. Hence, this work focuses on high performance robust SIFT based feature extraction of various earthquake affected areas from high resolution imagery and subsequent classification of using Support Vector Machines (SVM). The high performance computing frameowrk consists of Tesla C2075 Graphics processing unit (GPU) with 448 cores. Results obtained from GPU implementation is shows significant gains in computational time over CPU based approach.

Recently an emerging area that focuses on mining data across domains is Cross Domain Similarity Mining and analogical reasoning for data analysis. Geospatial domain is highly interdisciplinary with methods that incorporate multiple other domains; this research focuses on analogy reasoning using geo-ontology matching. Here a framework is proposed for analogy reasoning. Using an example of road networks, it is explained how analogy can be formulated and transferable and learnable patterns are generated.

With advances in sensing systems attempts are continuously being made to design Internet of Things (IoT) based interoperable sensing systems. The important issues (water, pest/disease, nutrient management, etc.) pertaining to crop-weather-soil continuum can be addressed through high resolution monitoring of agro-meteorological parameters. Presently the designed sensing systems have syntactic and semantic heterogeneity and face underlying limitations for achieving interoperability among these distributed sensing systems. In this study an attempt has been made to develop KrishiSense. A semantically aware web enabled wireless sensing system for precision agriculture applications. Through integration of Open Geospatial Consortium (OGC) specified Sensor Web Enablement (SWE) standards on sensing system has enabled the interoperability between different standardized sensing systems. KrishiSense acts as interconnection between multiple users (researchers/scientists, farmers and extension community) through multiple protocols and distributed web connected platforms, thus facilitating human participatory sensing.

Disaster image information mining is one of the crucial aspects in remote sensing applications. In a post disaster situation, to build learning model, new training samples are required, which are difficult to obtain. With the available pre-disaster data, the traditional algorithms cannot generalize well on the post-event situation for classification because, the data distributions are different. The proposed approach addresses this problem by domain adaptation to classify a post-disaster event by leveraging distribution changes. In this way it can augment the paucity in ground truth by using the prior information available to build the model.

Cloud detection and removal forms an important need for change detection studies. A small amount of cloud cover may misinterpret the crucial information in disaster management applications. Although several cloud detection techniques exist, there is a critical need to apply these techniques in real time and obtain the cloud free images quickly to support real time decisions.

Geospatial information is different than conventional information. Harmonization is needed for interoperability and seamless access to data. Ontology matching is an emerging solution to achieve this harmonization. The input data of the Geospatial ontologies vary from the conventional ontologies and hence it is conceptualized in a different manner. There are two major obstacles for geoinformation fusion: heterogeneity and uncertainty. Heterogeneity is more prevalent and uncertainty is an unavoidable entity in geospatial domain. This paper explores a novel multi-objective algorithm for geospatial ontology matching. It uses Pareto ranking to sort the probable solution and derives the pareto front. This pareto front is used further to find the best match.

Near real time processing and clusters extraction from high-resolution satellite images of disaster affected area aids in monitoring and deployment of rescue activities. In this work the k-medoids clustering algorithm is analyzed for near real time applications. In general, due to the large size of satellite data, the computational time of traditional k-medoids is found to be very high. Hence in order to achieve the aim of near real time processing of such huge data we developed a parallel implementation of k-medoids (GPUPAM) and integrated with CLARA (Clustering for LARge Application), which is one of the variant of kmedoids. The implementation is performed on NVIDIA's Graphical Processing Unit (GPU). The performance improvements that were obtained is demonstrated by a GPU implementation on high resolution Haiti Earthquake QuickBird (2.4 meter resolution) data and compared with the traditional sequential implementation. The results show that the GPU implementation is found to achieve almost 96% performance improvement as compared to the sequential implementation.

Geographic data is often collected from independent sources and is usually heterogeneous in nature. Integration of these heterogeneous data sources is a crucial task. The recent developments in the semantic web domain have shown great potential to address the geospatial data integration issues. Ontology matching is seen as a solution for integration problems and has attracted wide attention. Geospatial domain is characterized by vagueness moreover semantic ambiguity leads to uncertainty in developing ontology and this is propagated to the ontology-matching phase. Hence, to resolve the uncertainty issues this study focuses on the adaptation of information theory based approaches for geospatial ontology matching.

In a developing country like India with its overgrowing population, increase in agriculture productivity and its monitoring is an important concern. Hence, environmental data in conjunction with crop information (e.g. soil moisture) are necessary for crop management and productivity enhancement. Enhanced monitoring and seamless information exchange in real time is only possible if data from these sensors are standardized. This also enables integration of data from several related areas of interest and facilitates interoperability. We propose Open Geospatial Consortium (OGC) standards-based Service-Oriented Architecture (SOA) for data integration. The Sensors Web Enablement (SWE) framework from OGC has been implemented for selected sensors belonging to the agricultural domain. OGC's Sensor Observation service (SOS) (part of SWE suite) is adapted for discovery and access of sensor observations (real-time or archived) and Sensor Event Service (SES) for sending alerts and notifications.

The available empirical remote sensing techniques for harmful algal bloom (HAB) detection are reliant on prior observations and thresholds. These techniques tend to give high false alarm rate, as they are limited in spatiotemporal contextual information and decision combination techniques. We propose a multistage learning based ensemble methodology addressing the above constraints for performance improvement of HAB detection. Machine learning-based spatiotemporal data mining approach, along with empirical relationships, is used for HAB detection in the first stage of the ensemble, to exploit the potential benefits of each individual detection technique. The decision outputs from these detection techniques are fused in the second stage using nonlinear modeling-based combination techniques unlike conventional weighted averages. The proposed ensemble methodology outperforms all of the individual members and gave a significant overall performance improvement up to 0.8632 kappa accuracy. The performance is evaluated over tenfold cross validation average and compared against various ensemble methods and combination techniques.

The recent explosion of data from various Earth observation (EO) systems requires new ways to rapidly harness the information and synthesize it for decision making. Currently several image information mining (IIM) systems have some form of supervised statistical learning models that relate the image content to the various semantic classes. However, this kind of approach is constrained by the paucity of training information in several EO domains due to limited ground truth. Although, semi-supervised learning methods alleviate this problem to a certain extent by using unlabelled data from various spatial databases, however these methods require that the training data and future unseen data should conform to the same statistical distribution and feature space. To overcome this problem a more recent approach is focused on using small amounts of labeled information from closely related or similar learning task and somehow adapt that information in developing new semantic models. The above methodology called transfer learning can be applied in several processes of supervised and unsupervised learning. In this paper, we propose Transfer learning methods for IIM and discuss various techniques and their implications for content-based retrieval in the EO domain. Specifically, we explore transfer learning application in a rapid disaster response scenarios during coastal events. The adopted methodology for knowledge transfer is based on harnessing prior knowledge from similar concepts to learn new ones and uses a modified weighted least squares support vector machine (SVM).

The present study discusses about the new textural feature extraction, its improvement and a comprehensive analysis of our previous Machine Learning based Spatio-Temporal (STML-HAB) Data Mining approach for HAB detection mentioned in Ref. [2]. This study is an elaborative analysis extending our first results presented in Ref. [2]. The additional Wavelet and GLCM textural features helped in improving the performance up to an accuracy of 0.9259 'K' using SeaWiFS sensor data. This is a significant improvement of almost 17% compared to our first results with an accuracy of (0.7513 'K').

The Combinational algorithm in the ensemble approach plays a key role towards the performance. The standard majority voting, weighted average and probabilistic averaged weight could not tune well the decisions of the multi-classifiers to the class label. We propose the modeling of the multi-classifier decisions to the output variable using Probabilistic Neural Networks as the combinational algorithm. This proposed implementation of combinational algorithm gave a significant performance improvement against the standard combiners.

In a disaster, there is a need for rapid image-information retrieval in real or near real time from vast amounts of data coming from multiple remote-sensing sensors. In general, image information mining (IIM) approaches produce enormous amounts of features that are computationally expensive and inefficient to process before the actual information discovery takes place. Also, it is complicated because the combination of the features has little relevance to the hypothesis space. Hence, selecting a relevant subset of features is necessary to overcome these problems and to provide an efficient representation of the target class. In this letter, we propose feature selection and feature transformations based on a wrapper-based genetic algorithm approach. A support vector machine classification is applied for generating predictive models for those land-cover classes that are important in a coastal disaster event. The proposed system, rapid IIM, is a region-based approach where, in lieu of the prevalent pixel-based methods, it localizes interesting zones and enables rapid querying. Results from this study indicate that selecting relevant feature subsets increases the rate of correctly identifying a semantic class and also enables this process with less number of features.

Image transformation is required for color-texture image segmentation. Various techniques are available for the transformation along the spatial and spectral axes. For instance, the HSV-wavelet technique is shown to be very effective for image information mining in remote-sensing applications. However, the HSV transformation approach uses only three spectral bands at a time. In this letter, a new feature set, obtained by combining independent component analysis and wavelet transformation for image information mining in geospatial data, is presented. Experimental results show the effectiveness of the presented method for image information mining in Earth observation data archives.

A new generation of mobile device users is coming of age in the next decade. These users can explore the mobile internet with its new features, services, and applications. Recently, an application platform like the Google's Android mobile platform has revolutionized open applications development for the mobile platform. As increasing number of companies expose their services as web services, enabling flexible mobile access to distributed Web resources is a very relevant challenge. However, the current web is a collection of human readable pages that are unintelligible to computer programs. Semantic Web and Web services have the potential of overcoming this limitation. Semantic Web technology and the advent of universal and mobile access to Internet services, provides additional features like knowledge-based, location or context aware information. For this, a standard ontology called ontology Web language for services (OWL-S) is employed. The vision is to automatically discover services like sensor Web service, geospatial information service, etc from mobile. In this work, we apply the above methods to the coastal sensor Web.

Ontologies are widely used as a means for solving the information heterogeneity problems on the Web because of their capability to provide explicit meaning to the information. In recent years, the growing need to resolve the ambiguities and integrate vocabularies between heterogeneous systems within a domain of interest led to the rapid development of ontologies by different organizations. These ontologies designed for a particular task could be a unique representation of their project needs. Hence, there arises a need to align heterogeneous ontologies to facilitate meaningful knowledge interchange between various sources. Thus, ontology mapping has become the key point to enable semantic interoperability between different representations within a domain. This paper proposes a new instance-based algorithm to automate ontology mapping in ocean sensor networks, whose data are highly heterogeneous in syntax, structure and semantics.

Harmful Algal Blooms (HABs) pose an enormous threat to the U.S. marine habitation and economy in the coastal waters. Federal and state coastal administrators have been working in devising a state-of-the-art monitoring and forecasting system for these HAB events. These modernized HAB systems provide useful and forewarning information to a varied user community. However, the lack of standardization in the data exchange mechanism with the current available systems causes an impediment to the wide area coastal observation and management. Hence, there is a need for the system to adapt the services oriented architecture and the OGC (Open Geospatial Consortium) sensor web enablement framework. We propose a HAB monitoring system by adopting the standardized OGC sensor web and using machine learning approaches for the detection of HAB events in the region of Gulf of Mexico. Various feature extraction techniques have been used in obtaining features of both HAB and Non-HAB data. Kernel based Support vector machines have been used as a classifier in the detection of HAB's. The performance of this approach is analyzed by accuracy measures like Kappa Coefficient, N-fold cross validation average and Confusion Matrix on a considerable test data.

In this paper we present details of a proposal to develop standards for the processes involved in image information mining (IIM) and also to initiate a discussion within the IIM community to evolve and develop specifications. The proposal is based on the current Services Oriented Architectures (SOA's) which provides loosely coupled services that enable cross domain information integration and querying. Web services decouple objects that are platform specific and facilitate interactions among platform independent objects, which are able to access data from anywhere on the Web. They rely on loose, rather than tight coupling among the web components which enables a flexible and dynamic interchange in open, distributed web environments. The Open Geospatial Consortium (OGC) provides a platform for government organizations, academia, and industry to come to a consensus and standardization of geospatial technologies. This paper proposes that image information mining systems need standardization in terms of OGC specifications and in describing the IIM framework in an OGC perspective. This would facilitate interoperability with several existing OGC web services and foster the clear separation of the business logic layer and presentation layer.

In this paper we propose a framework that focuses on the need for rapid image information mining in a coastal disaster event where it is necessary to explore vast amounts of data from multiple remote sensing sensors in real or near real time. The proposed system; Rapid Image Information Mining (RIIM) is a region based approach where inlieuof prevalent pixel based methods it localizes interesting zones and extracts information from them that are stored in a knowledge base. A set of primitive features are extracted from the regions, whose relevance for a particular land cover class or a combination of classes is then assessed based on a wrapper based genetic algorithm (GA) approach. In this, we use an induction algorithm along with the GA to arrive at an optimal set of features. We investigate feature selection and feature generation using this wrapper approach. A support vector machines based classification is applied for generating predictive models for those land cover classes that are important in coastal disaster events. In RUM, searching for a particular land cover type (e.g. flooded agriculture) is based on the actual meaning and content of it in the image instead of just the metadata.

This paper presents an image segmentation method using contourlets for spatial transformation. Independent component analysis (ICA) is used to obtain features that are independent and uncorrelated. Feature reduction is also performed during the preprocessing stage of the ICA. A kernel- based approach for clustering the dataset will eliminate the need to calculate cross-correlation energies explicitly. Instead of performing clustering over the whole image, two stages of kernel- based clustering help in reducing the computation complexity of the segmentation method. The segmentation method is applied to LandSat 7 ETM+ imagery. Results show a promising use of the presented approach for image information mining within a semantic framework. These primitive features can subsequently used for object identification.

Recently, wavelet-based methods have been efficiently used for segmentation and primitive feature extraction to expedite the image-retrieval process of semantic-enabled frameworks for image information mining from geospatial data archives. However, the use of wavelets may introduce aliasing effects due to subband decimation at a certain decomposition level. This paper addresses the issue of selecting a suitable wavelet decomposition level, and a systematic selection process is developed. To validate the applicability of this method, a synthetic image is generated to qualitatively and quantitatively assess the performance. In addition, results for a Landsat-7 Enhanced Thematic Mapper Plus imagery archive are illustrated, and the F-measure is used to assess the feasibility of this method for the retrieval of different classes

The Global Earth Observation System of Systems (GEOSS) is a distributed system of systems built on current international cooperation efforts among existing Earth observing and processing systems. The goal is to formulate an end-to-end process that enables the collection and distribution of accurate, reliable Earth Observation data, information, products, and services to both suppliers and consumers worldwide. Earth Observations (EO) are obtained from a multitude of sources and requires tremendous efforts and coordination among different agencies and user groups to come to a shared understanding on a set of concepts involved in a domain. Semantic metadata plays a crucial role in resolving the differences in meaning, interpretation, usage of the same or related data. Also the knowledge about the geopolitical background of the originating datasets could be encoded in the metadata that would address the diversity on a global scale. In distributed environments like GEOSS modularization is inevitable. In this paper we propose a framework for modular ontologies based knowledge management approach for GEOSS in which we explore approaches on formulating smaller interconnected ontologies. This analysis is exercised in a coastal zone domain.

Reflectance pattern and spatial pattern characterize the geospatial data. Current semantic-enabled framework retrieval system extract primitive features based on color, texture (Spatial Gray Level Dependency - SGLD matrices), and shape from the segmented homogenous region. This system can use only three bands (true color or false color) at a time to capture color information as it converts RGB space into HSV space. Thus it fails to capture the complete reflectance pattern, an important characteristic of geospatial data. This paper describes a new method to perform image segmentation using the features obtained by coalescing of Independent Component Analysis and Wavelet transform, which are later on used for the region-based retrieval in the earth observation data archives. Experimental results show effectiveness of the proposed method for image information mining in Earth observation data archives.

Ontologies are widely recommended as a means of rectifying semantic heterogeneity. The advantage of using ontologies is that they can provide a conceptual schema regardless of a data set’s format, structure, or size MSU is developing an ontological framework for resolving semantic heterogeneity problems in coastal zone data. This type of framework will provide the capability to (a) link the users to the knowledge, making integrated visualizations available; (b) provide search and query answering facilities; and (c) gather information at different levels of granularity, from the subcategory to the specific data level. Issues related to coupling such a system to models will also be discussed.

Earth observation data have increased significantly over the last decades with satellites collecting and transmitting to Earth receiving stations in excess of 3 TB of data a day. This data acquisition rate is a major challenge to the existing data exploitation and dissemination approaches. The lack of content- and semantic-based interactive information searching and retrieval capabilities from the image archives is an impediment to the use of the data. In this paper, we describe a framework we have developed [Intelligent Interactive Image Knowledge Retrieval (I/sup 3/KR)] that is built around a concept-based model using domain-dependant ontologies. In this framework, the basic concepts of the domain are identified first and generalized later, depending upon the level of reasoning required for executing a particular query. We employ an unsupervised segmentation algorithm to extract homogeneous regions and calculate primitive descriptors for each region based on color, texture, and shape. We initially perform an unsupervised classification by means of a kernel principal components analysis method, which extracts components of features that are nonlinearly related to the input variables, followed by a support vector machine classification to generate models for the object classes. The assignment of concepts in the ontology to the objects is achieved automatically by the integration of a description logics-based inference mechanism, which processes the interrelationships between the properties held in the specific concepts of the domain ontology. The framework is exercised in a coastal zone domain.

The earth observation data has increased significantly over the last decades; NASA has 18 Earth observation satellites on orbit earning 80 sensors, as of April 2003. About 3 terabytes of data are collected daily and transmitted to Earth receiving stations. The data exploitation and dissemination methods have not kept pace with the huge data acquisition rate. The products distributed by the agencies are often not in a readily usable form by the nonscience community, and need further processing at the user level. The lack of content and semantic based interactive information searching and retrieval capabilities from the archives is another important issue to be addressed in this context. We propose a framework based on a concept-based model using domain-dependant ontologies where the basic concepts of the domain are identified first and generalized later depending upon the level of reasoning required for executing a particular query. We employ an unsupervised segmentation algorithm to extract homogeneous regions and calculate primitive descriptors for each region based on color, texture and shape. The primitive descriptors are described quantitatively by middle level object ontology. The learning phase is applied at this stage. It associates the middle level descriptors to the concepts in the higher-level ontology by means of a nonlinear support vector machine (SVM) method. These associations are grouped into models specific to a semantic class and used for querying. Also interactive querying is provided by means of a region based relevance feedback method. A methodology to execute complex queries by the integration of an inference engine is discussed. We also intend to extend the system to carry out data exploratory tasks in a peer-to-peer environment.

The plethora of new satellites that collect vast amounts of imagery in specific bands and multiple view angles has given rise to the enormous task of managing and storing large quantities of data and also to extract meaningful information from it. Efforts are underway to develop methodologies that will query the images by content, provide efficient visualization and perform data mining. While the continuation of deploying such state of the art satellites in the future weighs heavily on the logistics of the space agencies, the user community will be in a quandary to look for data that will enable them to test algorithms and methodologies. This paper describes the development of a virtual remote sensing paradigm of simulating atmosphere, scene, and sensor parameters. This is achieved by the modification of existing models available in these areas and coupling them in a software environment developed in Interactive Data Language (IDL). A graphical user interface enables the user to interactively select different land cover types (e.g. water bodies, agricultural fields, soils etc) from a predefined menu and create a virtual scene. A scene is defined by determining its size x columns by y rows, where each location (x, y) is a square scene cell determined by the resolution corresponding to the sensor selected by the user and is assigned to one of the different land cover classes. A Radiative transfer code, Streamer, has been used to model the atmospheric effects. A number of available kernel models have been incorporated in the software to determine bidirectional reflectance corresponding to a given surface type. Field collected BRDF data by a goniometer has been used to perform model inversion and obtain the model parameters.