China’s first Mars mission (Tianwen-1) successfully touched down in the Utopia Planitia of Mars with a rover subsurface penetrating radar (RoPeR) carried for exploring the regolith dielectric properties. Hyperbolic fitting is a conventional method to infer the subsurface material relative permittivity from ground penetrating radar data (GPR). However, it is difficult to directly extract valid hyperbolas from the RoPeR data. Inspired by the recently developed deep learning-based geophysical inversion method to estimate of the subsurface wave velocities through GPR data, an improved deep learning architecture is proposed to infer the Martian regolith relative permittivity from the RoPeR data, with self-attention and cascade modules are introduced into the network. The improved cascade and self-attention modules can improve the inversion efficiency and mitigate the scatter-diffraction effect of the predicted results. The inverted relative permittivity from the first 60 ns of the RoPeR data demonstrates an approximate line with a mean value of 4.73 in the regolith of interest. The very limited fluctuation of relative permittivity implies that no explicit stratification existing in the investigated regolith, agreeing with the previous studies.

This study reveals that the magnetotelluric (MT) method can detect the responses of the ionospheric E-region to the annual solar eclipse on June 21, 2020. The MT method not only detected a significant reduction in the solar quiet current by approximately 3.3 h over Lijiang during the obscuration but also, for the first time, observed waves with periods ranging from a few to dozens of minutes in the electric field due to the eclipse. The evolution of electric and magnetic fields suggests nonstationary changes in the ionospheric E-region structures during and after the eclipse.

Under the broad definition of multimodal data, data presenting different views and complementary information are classified as multimodal. Two-dimensional magnetotelluric (MT) responses, including apparent resistivity and phase of two polarization modes, reflect differing physical properties and offer complementary insights into the subsurface media. Traditional deep learning (DL) approaches often struggle to capture and integrate these complementary features effectively for accurate inversion. In this study, we treat$\rho _{S}^{\text {TE}} $,$\varphi ^{\text {TE}}$,$\rho _{S}^{\text {TM}}$, and$\varphi ^{\text {TM}}$of 2-D MT as multimodal data and introduce a data fusion method of multimodal DL (MDL), which enhances the accuracy of MT inversion by employing a multimodal U-net model to integrate various MT response features effectively. In detail, each MT response is processed in a different encoder to exploit its unique information better. It is densely connected within each encoder and across different encoders, facilitating the fusion of MT data across depths and response modalities. Our method maximizes complementary information from multiple response modalities, resulting in a more precise depiction of nonlinear processes in MT inversion. First, 2-D Gaussian random fields (GRFs) simulate the resistivity model. Then, the multimodal U-net is introduced and improved as the MT inversion framework and compared with the conventional U-net. Moreover, the anti-noise ability and generalization of the multimodal U-net are tested by introducing varying noise levels into the MT responses. Finally, we validate our proposed method with MT field data from the Yanggao area in the Datong Basin, Shanxi Province, China, showing that the performance of our model surpasses conventional U-net and traditional nonlinear conjugate gradient (NCG) inversion methods.

The subsurface conductivity of geological media is often anisotropic, making three-dimensional (3D) anisotropic inversion of controlled-source electromagnetic (CSEM) essential for resolving complex geologic settings. However, compared to isotropic inversion, anisotropic inversion involves a substantially greater number of model parameters, increasing the severity of the non-uniqueness problem and enhancing interpretation complexity for large-scale field data. To address these challenges, we present an innovative anisotropic inversion approach that incorporates a Gramian-based constraint, which promotes similarity between horizontal and vertical conductivity models without relying on a prior information. We formulate the inverse problem within a Gauss-Newton framework and employ the finite element method on unstructured grids, leveraging parallel direct solvers for computational efficiency. Synthetic tests on complex anisotropic land and marine CSEM models show that the Gramian-constrained inversion significantly suppresses spurious anomalies and improves reliability compared to conventional anisotropic inversion. Application to field CSEM data from the Huaniushan Pb-Zn mining area in Gansu Province, China, demonstrates high consistency with real geology and drilling results. These findings highlight the proposed approach as a computationally efficient, robust, with direct relevance to hydrocarbon, mineral, and geothermal exploration.

The capturing subsurface structure through geophysical measurements can gain a more comprehensive understanding of groundwater distribution. While geophysical electromagnetic methods yield subsurface resistivity data, converting this into hydrological information is not straightforward. Well-logging offers insights into rock strata vertically but lacks spatial detail on large-scale lithological variations. Consequently, merging geophysical and well-logging data for extensive hydrogeological modeling has emerged as a crucial research area. In this study, we introduce convolutional neural networks and bi-directional long short-term memory (CNNs-BiLSTM) network to process massive towed transient electromagnetic (tTEM) datasets. Our network incorporates the depth-of-investigation (DOI) and smooth constraints for effective tTEM data inversion. We further validate the network’s effectiveness and generalization capacity using synthetic models and real tTEM data from Switzerland’s Aare Valley region. Furthermore, by combining the limited well-logging data, we establish a spatial clay content distribution model using an optimal inversion interpolation method. Leveraging this lithology model, we employ the groundwater modeling system (GMS) platform to determine regional groundwater levels. Our numerical simulation aligns closely with results obtained via the top of the saturated zone (TSZ) method and exhibits strong agreement with observed water table data, affirming the reliability of our comprehensive hydrogeological model. Our proposed method and workflow present an innovative approach to effective hydrological modeling utilizing large-scale geophysical electromagnetic data.

The resistivity distribution obtained from traditional inversion methods for magnetotelluric (MT) data often lacks clarity, making it difficult to delineate boundaries between host media and anomalous targets. To address this, we developed a novel 3-D MT inversion approach based on the multinary transformation of model parameters. This method transforms the model resistivity distribution into a desired step-function-like form, enabling explicit identification of interfaces between geological units. The sharpness of the recovered resistivity model is controlled by the standard deviation of the multinary transformation’s error function, and an adaptive technique is introduced to adjust this parameter during the inversion process to account for deviations between true and discrete values in the multinary space. The inversion problem is solved using a data-space Gauss-Newton approach, which enhances memory efficiency and convergence speed. Additionally, unstructured tetrahedral meshes are utilized to accurately model rugged topography and complex geoelectric structures. Synthetic model studies demonstrate the superiority of the adaptive multinary inversion over conventional maximum smoothness inversion and fixed standard deviation multinary inversion. Finally, the method is applied to image subsurface resistivity in the northwest Geysers geothermal field in California, USA, showcasing its effectiveness.

Lane-level high-definition map (HDMap) automatic construction from aerial imagery can significantly enhance the efficiency of HDMap production. However, some problems limit the application of aerial imagery in HDMap construction. First, the bird’s eye view of aerial imagery often causes occlusion, which leads to interrupted map element predictions. In addition, existing methods for aerial imagery-based HDMap construction can only construct road-level map and the postprocessing is complicated. To solve these problems, we propose AerialHDMapper, an end-to-end framework that predicts lane-level vectorized map elements directly requiring no postprocessing. The framework improves prediction accuracy in occluded areas through the introduction of the directional attention (DirAttn) module. AerialHDMapper is mainly composed of two parts: the image encoder and the polyline generator. The image encoder integrates DirAttn modules, which leverage information along and perpendicular to the direction of a map element, thereby improving the representation of the current position. The polyline generator employs a modified transformer architecture to identify long-sequence irregular map elements. Experimental results demonstrate that AerialHDMapper outperforms existing methods on both the CARLA Simulator Dataset and the Aerial Argoverse2 Dataset we collected. Extensive ablation experiments further reveal that our approach exhibits strong robustness and generalization capabilities.

Underwater instance segmentation plays a pivotal role in various applications. Among them, coral instance segmentation is of great significance in the fields of marine biology and environmental monitoring, and is crucial for comprehensive understanding of coral reef ecosystems. Traditional methods for underwater instance segmentation predominantly rely on RGB images. However, the complex morphology of corals and strong background interference often result in poor segmentation outcomes. To tackle these problems, this study presents a novel multimodal instance segmentation method, termed UMIS-YOLO, which is grounded in the YOLO architecture. UMIS-YOLO incorporates a dual backbone network design that substantially enhances the feature extraction capabilities for both RGB images and depth images, thereby improving the effectiveness of instance segmentation. At the same time, we propose two innovative plug-and-play modules: the Frequency Domain Feature Enhancement Fusion (FDFEF) module and the Residual Feature Fusion (RFF) module. The FDFEF module leverages Fourier transform to enhance the features of both modalities in the frequency domain, employing learnable weights to enable the complementary integration of amplitude and phase information. While the RFF module utilizes a residual learning strategy to efficiently merge low-level and high-level features prior to the segmentation head, thereby improving pixel-level segmentation accuracy. Additionally, we introduce a challenging high-resolution dataset, UMIS-Coral, which comprises RGB images and depth images captured in complex coral environments. Meanwhile, we expand the depth images for the UIIS dataset to further verify the effectiveness of UMIS-YOLO. The experimental results indicate that the UMIS-YOLO model achieved mAP50 and mAP75 improvements of 2.3 and 3.0 on the UMIS-Coral dataset, as well as 3.9 and 2.8 on the UIIS dataset, respectively. Furthermore, the model is characterized by its lightweight architecture and rapid segmentation capabilities. The source code and the dataset are publicly accessible at https://github.com/zhangsanhulk/UMIS-YOLO.

Building vector maps play an essential role in many remote sensing (RS) applications, thereby boosting the deep learning (DL)-based automatic building vector extraction methods. These approaches have achieved pleasant overall accuracy, but their predict-style framework struggles with perceiving subtle details within a tiny area, such as corners and adjacent walls. In this study, we introduce a denoising diffusion framework called DiffVector to generate representations for direct building vector extraction from the RS images. First, we develop a hierarchical diffusion transformer (HiDiT) to conditionally generate robust representations for detecting nodes and extracting corresponding features. The conditions of HiDiT are multilevel boundary attentive maps encoded from input RS images through a topology-concentrated Swin Transformer (TCSwin). Subsequently, an edge-biased graph diffusion transformer (EGDiT) takes extracted node features as conditions to produce new visual descriptors for the adjacency matrix prediction. In EGDiT, we replace the standard self-attention (SA) operation with an edge-biased attention (EBA) to inject edge information for training stabilization. Furthermore, given typical challenges of training difficulty and weak perceptive ability in convectional diffusion paradigms, we conduct an isomorphic training strategy (ITS), ensuring that the training procedures of both HiDiT and EGDiT precisely mirror the inference phase. Quantitative and qualitative experiments have evidently demonstrated that DiffVector can achieve competitive performance compared with existing modern approaches, especially in the metrics assessing topology quality.

High-quality image segmentation is fundamental to pixel-level geospatial analysis in remote sensing, necessitating robust segmentation quality assessment (SQA), particularly in unsupervised settings lacking ground truth. Although recent deep learning (DL) based unsupervised SQA methods show potential, they often suffer from coarse evaluation granularity, incomplete assessments, and poor transferability. To overcome these limitations, this paper introduces Panoramic Quality Mapping (PQM) as a new paradigm for comprehensive, pixel-wise SQA, and presents SegAssess, a novel deep learning framework realizing this approach. SegAssess distinctively formulates SQA as a fine-grained, four-class panoramic segmentation task, classifying pixels within a segmentation mask under evaluation into true positive (TP), false positive (FP), true negative (TN), and false negative (FN) categories, thereby generating a complete quality map. Leveraging an enhanced Segment Anything Model (SAM) architecture, SegAssess uniquely employs the input mask as a prompt for effective feature integration via cross-attention. Key innovations include an Edge Guided Compaction (EGC) branch with an Aggregated Semantic Filter (ASF) module to refine predictions near challenging object edges, and an Augmented Mixup Sampling (AMS) training strategy integrating multi-source masks to significantly boost cross-domain robustness and zero-shot transferability. Comprehensive experiments demonstrate that SegAssess achieves state-of-the-art (SOTA) performance and exhibits remarkable zero-shot transferability to unseen masks. The code is available at https://github.com/Yangbn97/SegAssess.

Interactive segmentation (IS) using minimal prompts like points and bounding boxes facilitates rapid image annotation, which is crucial for enhancing data-driven deep learning methods. Traditional IS methods, however, process only one target per interaction, leading to inefficiency when annotating multiple identical-class objects in remote sensing imagery (RSI). To address this issue, we present a new task—identical-class object detection (ICOD) for rapid IS in RSI. This task aims to only identify and detect all identical-class targets within an image, guided by a specific category target in the image with its mask. For this task, we propose an ICOD network (ICODet) with a two-stage object detection framework, which consists of a backbone, feature similarity analysis module (S3QFM), and an identical-class object detector. In particular, the S3QFM analyzes feature similarities from images and support objects at both feature-space and semantic levels, generating similarity maps. These maps are processed by a region proposal network (RPN) to extract target-level features, which are then refined through a simple feature comparison module and classified to precisely identify identical-class targets. To evaluate the effectiveness of this method, we construct two datasets for the ICOD task: one containing a diverse set of buildings and another containing multicategory RSI objects. Experimental results show that our method outperforms the compared methods on both datasets. This research introduces a new method for rapid IS of RSI and advances the development of fast interaction modes, offering significant practical value for data production and fundamental applications in the remote sensing community.

Gravity and magnetic anomalies contain abundant geological information. However, redundant information complicates the study of exploration targets. Existing methods primarily rely on exploiting spectral differences between shallow and deep sources to separate anomalies of different depths. Nevertheless, spectral overlap limits these conventional methods to separating anomalies caused by significantly different depth sources. To reduce effects due to spectral overlap, we propose a novel method for potential field separation. This method capitalizes on the sparsity of gravity and magnetic data in both singular spectrum and model spaces and employs a single-layer equivalent source to represent anomalies induced by target sources. The anomalies caused by sources with different depths can be separated. After sparsely approximating single-layer equivalent sources, we obtain the local anomalies caused by sources within the same layer. Synthetic model experiments demonstrate that the proposed method achieves high separation accuracy, particularly with respect to effectively separating anomalies induced by models with small depth differences. In addition, when comparing the noise resistance of low-rank methods with existing potential field separation methods using synthetic data, the results show that low-rank methods can extract effective signals from signals contaminated by sparse noise and periodic noise. We then apply this method to extract local gravity anomalies caused by intrusive rocks in the Nanling region and effectively identify gravity anomalies associated with various intrusive rocks. This method facilitates the separation of gravity and magnetic anomalies originating from sources at both different and similar depths, thereby expanding the applicability of separation techniques and enhancing the resolution of gravity and magnetic detection.

In this research, we introduce the enhanced automated quality assessment network (IBS-AQSNet), an innovative solution for assessing the quality of interactive building segmentation within high-resolution remote sensing imagery. This is a new challenge in segmentation quality assessment, and our proposed IBS-AQSNet allievate this by identifying missed and mistaken segment areas. First of all, to acquire robust image features, our method combines a robust, pre-trained backbone with a lightweight counterpart for comprehensive feature extraction from imagery and segmentation results. These features are subsequently fused using simple convolution layers with residual connections. Additionally, IBS-AQSNet incorporates a multi-scale differential quality assessment decoder, proficient in pinpointing areas where segmentation result is either missed or mistaken. Experiments on a newly-built EVLab-BGZ dataset, which includes over 39,198 buildings, demonstrate the superiority of the proposed method in automating segmentation quality assessment, thereby setting a new benchmark in the field.

The frequency of the free induction decay (FID) signal induced from an Overhauser magnetometer sensor is proportional to the magnetic field to be measured. Due to the low initial signal-to-noise ratio (SNR), sensor tuning is necessary to suppress the noise and improve the frequency estimation accuracy. To improve the tuning performance in complex strong-disturbance environments, this study introduces a novel method using higher-order singular value tensor decomposition (HOSVTD) and Fourier synchrosqueezing transform (FSST), namely HOSVTD-FSST. First, multiple FID signals are obtained using an equal delay multichannel acquisition strategy to establish a deeper, more intrinsic correlation attribute. Second, matrix segmentation is applied to construct the signals into a higher-order tensor for singular value computation, and the CANDECOMP/PARAFAC (CP) decomposition is fused to obtain a low-noise FID. Third, the FSST is employed to analyze the low-noise signal to extract the time-frequency ridges to capture the tuning frequency. Finally, the HOSVTD-FSST is compared with numerous commonly used methods. The experimental results demonstrate that under the presence of spike noise and with the SNR less than −20 dB, the frequency tuning deviations of the commonly used methods are up to 100 Hz, while that of the HOSVTD-FSST is within 5 Hz, which verifies that the HOSVTD-FSST can significantly enhance the sensor tuning accuracy in complex strong-disturbance conditions.

Semantic segmentation of airborne laser scanning (ALS) point clouds is a valuable yet challenging task in remote sensing. When processing large-scale ALS scenes, it is necessary to partition them into smaller blocks for ease of handling. However, this partitioning introduces a challenge in capturing the ample spatial context within each block to adequately recognize the objects with a significant spatial span. This limitation becomes particularly pronounced when relying solely on the 3D representations as the input of nerual networks. To incorporate sufficient contextual information in ALS data semantic segmentation, we propose a multi-modal-based segmentation framework called projection-based context embedding (PCE) in this study. PCE effectively combines the advantages of 2D image and 3D point-voxel representations, which are the computational efficiency and the representation capability for fine-grained 3D geometries. The 2D projection is used to encode a large-scale semantic context, which is computationally expensive to be obtained using only pure 3D representation. Simultaneously, the sparse-point-voxel convolution (SPVConv) is employed to focus on learning 3D features from a small block of points centered on the large-scale context. Finally, to fully exploit the power of each modality, the embedding disentangling (ED) strategy is proposed additionally to combine the context embedding from the 2D image with 3D features for the final prediction. We demonstrate the state-of-the-art performance of PCE through extensive experiments on public large-scale ALS point cloud datasets.

Accurate estimation of the Earth’s interior temperature is essential for solving fundamental scientific and applied geothermal problems. Currently, there is no universal method for determining deep temperature fields; however, such a method may be based on resistivity, a temperature-dependent proxy parameter. We propose an electromagnetic (EM) geothermometer based on the temperature–pressure coupled resistivity model (TPCRM). This geothermometer can accurately determine the relationship between the normalized resistivity, temperature, and pressure in deep formations based on well-logging, gravity, and EM data, thus allowing to visualize the temperature distribution. The TPCRM is utilized to predict the subsurface temperature in the Xiong’an New Area and shows an accuracy of 76.35%–96.58%. Sensitivity analysis of the critical variables of the TPCRM reveals that the TPCRM relatively weakly depends on the number of constraining boreholes and that the optimization of the subdivision spacing of the well-logging data can significantly improve temperature prediction accuracy. In addition, the effect of the spacing of inverted resistivity normalization grid nodes on the temperature prediction accuracy is relatively weak because the TPCRM considers the factor of the overburden pressure. The TPCRM is a promising tool for studying thermal genetic mechanisms, as well as fine evaluation of geothermal resources for their large-scale and efficient development and utilization.

We present a multi-dimensional inversion methodology for loop source time-domain electromagnetic data. The developed algorithm is a robust, efficient, and user-oriented tool for the multi-dimensional inversion of typical loop source time-domain electromagnetic configurations. A time-domain finite volume method and a direct solver are utilized for solving the 3D forward problem, while the iterative Gauss-Newton optimization method is implemented for the inversion kernel. The code is parallelized for calculating multiple sources simultaneously to accelerate the inversion. Based on different exploration tasks, we present three different inversion experiments on typical field scales for commonly used loop source TEM configurations. These examples verify the effectiveness and benchmark the developed 3D algorithm. Considering that TEM data is often gathered along profiles, we carried out an analysis using the 3D inversion algorithm for 2D data by adjusting the model roughness along the different domain directions, which seem to sufficiently constrain but not over-regularize the model along the strike direction to retrieve the true model structure. Besides using the vertical signal component for large scale moving and fixed loop configurations, the horizontal components are also included in the 3D inversion. The inversion algorithm is further verified with dense boat-towed central loop TEM data recorded on a volcanic lake on the Azores, Portugal. The reconstructed model is consistent along the profile line and the main features agree with the 1D stitched models. Moreover, the 3D model depicts additional reasonable features not visible in the 1D models.

In this study, we present an efficient and memory saving 3D inversion algorithm for interpreting controlled-source electromagnetic (CSEM) data using the total electric field formulation. To tackle CSEM problems involving complex geometries, we discretize the study domain for both forward and inversion problems using unstructured tetrahedral elements. Our inversion scheme combines the parallelized finite element (FE) method with the Gauss-Newton optimal strategy. Additionally, we transform the conventional model space inversion into data space inversion, significantly reducing the computation time and Random Access Memory (RAM) requirements during the inversion process. To begin, we validate the effectiveness and stability of the developed data space inversion algorithm by utilizing a synthetic land CSEM basin model and a synthetic marine CSEM model with bathymetry. These validation experiments further demonstrate that compared with the conventional model space inversion method, the computational efficiency of the data space inversion scheme is greatly improved and the memory required is significantly reduced. Furthermore, we apply the inversion method to survey CSEM data to demonstrate the practical applicability of the new inversion scheme.

Image segmentation using deep learning has become overwhelmingly widespread. However, routine model testing methods can encounter evaluation inconsistencies or bias, largely due to how accuracy metrics respond to variations in class share distribution. Here, we address the effects of class imbalance on model performance evaluation and demonstrate a refined approach that incorporates image classification efficacy (ICE) metrics within the context of semantic segmentation in remote sensing. This evaluation approach was applied in six segmentation experiments that involved multispectral and LiDAR data, single or multiple models tested with the same or different datasets, and binary and multiclass schemes. ICE metrics revealed unique aspects of model’s segmentation capabilities compared to precision, recall, F-score, and overall accuracy. By mitigating the class imbalance effect, per-class efficacy enables precise class-level optimization of segmentation models, while whole-class efficacy facilitates evaluating a model’s potential performance when adapted to new datasets. The suitability of the kappa coefficient, ROC-AUC, and PR-AUC for model evaluation under class imbalance was discussed in comparison with ICE metrics. This efficacy-enhanced model evaluation protocol can be implemented for deep learning model training and testing. The routine use of this evaluation approach will strengthen the dependability and applicability of segmentation tools in various fields.

Fourier neural operator (FNO) is a recently proposed data-driven scheme to approximate the implicit operators characterized by partial differential equations (PDEs) between functional spaces. The infinite-dimensional functional mapping from the parameter space to the state variable space enables us to solve parametric PDEs efficiently. To explore the potential of neural operator learning in geophysics exploration, we devise a transfer learning approach with the fine-tuning FNO backbone, termed transfer learning FNO (TL-FNO), to gain good generalization ability in solving frequency-domain wave equations at multiple source locations and frequencies. The baseline FNO model is initially trained at a single source location and frequency and then shared with the downstream tasks for seamlessly predicting the frequency-domain wavefields at different sources and frequencies. We conduct an in-depth analysis of the behavior of TL-FNO in diverse training settings, exploring dimensions such as data scale, training scale, and fine-tuning recipes. Our focus extends to understanding the scaling and transfer learning dynamics, as well as the generalization performance in out-of-distribution (OOD) scenarios. This comprehensive study aims to unveil the intricate relationships between these factors and the efficacy of TL-FNO across a range of conditions. Numerical examples demonstrate the notable superiority of the proposed TL-FNO over vanilla FNO in terms of accuracy and efficiency. We anticipate that the proposed TL-FNO is expected to be an efficient surrogate model to accelerate forward simulations in parametric wave equation inversion problems.

Recent learning-based models excel in point cloud registration for low-overlap scenes but falter in scenarios with minimal overlap. In this article, we propose a novel method to address the extreme case of low-overlap registration: non-overlapping point cloud registration. This scenario involves input point clouds that do not have overlapping regions but are adjacent to each other after registration. While the practical application value of non-overlapping point cloud registration remains to be explored, we believe that researching this issue contributes to enhancing the performance of registration in scenarios with extremely low overlap. Abandoning conventional overlapping region detection, we directly generate the registered source point cloud with SCREAM, a generative adversarial network (GAN). The generator incorporates information from the target point cloud into the source point cloud’s features and generates the registered source point cloud. To further align the generated results with the target point cloud, we propose a differentiable renderer that renders both the target and predicted point clouds into depth maps. These depth maps are then used as inputs to a discriminator to determine whether the generated results align with the target point cloud. Rigid transformation can be directly estimated from the correspondences between the source and the generated point clouds, bypassing the need for detecting overlapping regions, feature matching, and RANSAC steps found in previous methods. Extensive experiments demonstrate that SCREAM not only outperforms common overlapping point cloud registration scenarios but also achieves a registration success rate of 52.6% for the first time in non-overlapping scenes. We also constructed a new indoor scene registration dataset, 3DZeroMatch, specifically designed to explore non-overlapping registration problems. Our code and the dataset 3DZeroMatch are accessible athttps://github.com/xujiabo/SCREAM/.

We present a regularized inversion method for three-dimensional (3D) magnetotelluric (MT) data with axial anisotropic conductivities based on the edge-based finite element (FE) method. The Gauss–Newton (GN) approach is used to minimize the inversion objective function, including data misfit and regularization penalties, considering both structural complexity and anisotropic penalties. The most time-intensive task in the 3D MT inversion process is solving the large sparse system of linear equations. To speed up the inversion calculation, a hybrid direct–iterative solver combined with a block-diagonal preconditioner that has not yet been applied in anisotropic inversion is developed to accelerate the solutions for the sparse linear system resulting from forward modeling and sensitivity computations. In each GN iteration, a preconditioned conjugate gradient (PCG) method is adopted to overcome the difficulty of the sensitivity matrix storage for the anisotropic scene and obtain a model update without explicitly calculating and storing the sensitivity matrix. Before the inversion test, we use a model to demonstrate that the hybrid solver is computationally beneficial in terms of memory usage and time spent as compared to the direct solver. The good convergence properties and efficiency of the GN–PCG inversion scheme are demonstrated by two synthetic models and USArray data. The proposed inversion scheme can be an important supplement to existing anisotropic inversion algorithms and provide technical support for MT data interpretation.

Unexploded ordnances (UXOs) present a substantial and enduring threat to the human society in their vicinity. It is therefore imperative to explore effective and efficient methods for detecting UXOs. The coincident-loop transient electromagnetic (TEM) method is commonly used as a non-destructive technique for detecting UXOs. To gain a deeper understanding of TEM responses, a finite-element time domain (FETD) solver has been developed. This solver utilizes unstructured tetrahedral grids, enabling precise discretization of the real, complex shapes of UXOs. The FETD forward-modeling solver has been validated using the analytical method for the coincident-loop configuration. The numerical results from the landmine models show a significant impact of the metal shell thickness on the electromagnetic force (EMF), especially for a metal shell conductivity of 105S/m and shell thicknesses under 5 mm. In models incorporating both the projectile and fragments, the numerical findings indicate that a fragment, when positioned 1 m away from the projectile, can generate a similar EMF response to that of the projectile during the initial stages. Consequently, the clutter generated by fragments can have adverse effects on detecting the projectile. However, if a fragment is directly above or near the projectile, the fragment may not have a detrimental impact on detecting projectiles.

Rapid development in automatic vector extraction from remote sensing images has been witnessed in recent years. However, the vast majority of existing works concentrate on a specific target, fragile to category variety, and hardly achieve stable performance crossing different categories. In this work, we propose an innovative class-agnostic model, namely TopDiG, to directly extract topological directional graphs from remote sensing images and solve these issues. Firstly, TopDiG employs a topology-concentrated node detector (TCND) to detect nodes and obtain compact perception of topological components. Secondly, we propose a dynamic graph supervision (DGS) strategy to dynamically generate adjacency graph labels from unordered nodes. Finally, the directional graph (DiG) generator module is designed to construct topological directional graphs from predicted nodes. Experiments on the Inria, CrowdAI, GID, GF2 and Massachusetts datasets empirically demonstrate that TopDiG is class-agnostic and achieves competitive performance on all datasets.

The rapid processing, analysis, and mining of remote sensing big data using intelligent interpretation technology on remote sensing cloud computing platforms (RS-CCPs) have emerged as a new trend. However, existing RS-CCPs primarily focus on optimizing data storage and intelligent computing for common visual representation, overlooking key characteristics of remote sensing data such as large image size, large-scale change, multiple data channels, and geographic knowledge embedding. This oversight hinders computational efficiency and accuracy in remote sensing image interpretation. To address this, we have developed the LuoJia-AI platform, comprising the LuoJiaSET standard large-scale sample database and the dedicated deep learning framework, LuoJiaNET. This platform achieves state-of-the-art performance on five crucial remote sensing interpretation tasks: scene classification, object detection, land-use classification, change detection, and multi-view 3D reconstruction. LuoJia-AI bridges the gap between the sample database and the deep learning framework, exhibiting significant potential for high-precision remote sensing mapping applications.

Efficient three-dimensional magnetotelluric anisotropy forward modeling is one of the key research techniques used for inversion interpretation. We propose an improved multi-level down-sampling scheme to reduce the degrees of freedom of the stiffness matrix derived from the edge-based finite element method to improve the computational efficiency, saving on memory usage and calculation time for the forward modeling. Then, to further reduce the memory requirements and speed up the solution of the discretized electric system, we develop a multiple right-hand direct–iterative hybrid solver based on a block rational Krylov preconditioner. The solver we propose can further save computational costs and time based on the multi-level down-sampling scheme. Moreover, the convergence performance of the direct–iterative solver is less affected by the frequency, which solves the problem of slow convergence of the electric field control equation at low frequencies. We also use the high-level language Julia, which is easy to load into third-party packages, to ensure the stability and efficiency of the program. Finally, the validity and advantages of the two schemes are analyzed in detail using three examples. The example results show that the multiple right-hand direct–iterative hybrid solver and improved multi-level down-sampling can significantly reduce the computational memory and save computational time.

Automatic extraction of vector polygons of buildings from remotely sensed images is an important but difficult task. Recent existing methods based on deep learning usually adopt a multi-stage solution of semantic segmentation, contour detection, and polygon simplification. Such a long processing chain may lead to unreliable results as the boundary regularization and optimization processes are ultimately completed by utilizing low-level features, which ignores the potential of deep features in polygon generation. In this paper, we present an algorithm for directly extracting simplified polygons of buildings in remotely sensed images. The key of this task is the encoding of the polygon structure. PolyMapper [1] utilizes a recurrent neural network (RNN) to produce vertices of a polygon sequentially. Due to the limitation of RNN, this approach is unstable and difficult to deal with objects with complex shapes. In this work, we encode the polygon into a tensor representation and utilize a non-recurrent manner to recover the polygon structure. In our algorithm, two types of points are utilized, i.e., the corner point and the connecting point. Corner points are utilized to delineate the building outlines and form the vertices of the final polygon. Meanwhile, connecting points are sampled from the edges of the buildings for the assistance of the connection of the corner points. Furthermore, we predict the forward and backward directions of each keypoint in a polygon and propose a bidirectional tracing strategy for the polygon structure recovery. Our approach is simple, effective and robust. Experiments on public datasets demonstrate the superiority of the proposed algorithm. The code is made publicly available at https://github.com/sz94/bldvec.

Road topology extraction from satellite images, which has long been of interest, is an essential task in remote sensing. The graph representation of road networks is one of the most challenging aspects of road topology extraction. Most existing approaches cast road extraction as binary segmentation and then use postprocessing, such as skeletonization, to infer networks from pixelwise prediction. In our work, we believe that a road network can be represented by an undirected graph denoted as$G =$($V$,$E$), where$V$and$E$represent the set of road nodes and the set of edges between nodes, respectively. Thus, to construct the road topology, we propose NodeConnect, a new method of extracting nodes for a road network and inferring the connectivity between nodes. A convolutional neural network is jointly trained to predict the nodes and connectivity map for nodes, and the edges between nodes are inferred from the connectivity map. We compare our approach with several segmentation methods on the DeepGlobe and RoadTracer datasets. The experiments show that our approach achieves state-of-the-art performance in terms of pixel-based metrics and topological precision and recall.

The distribution of physical features in the Earth’s interior could be estimated by geophysical inversion from the acquired data at or above the surface. Inverse problems are generally considered as least-squares optimization issues in high-dimensional parameter space. Existing approaches are largely based on linear inversion methods, which are limited by the initial model. Nonlinear inversion methods, despite their significant ability in uncertainty quantification, still remain a formidable computational task. In this letter, a new gravity inversion approach is developed based on convolutional neural networks (CNNs). Although the training stage of this method is time-consuming, the actual prediction can be performed in only seconds. Thus, the high computational time of geophysical inversion can be considerably decreased once an appropriate network is constructed. The tests on synthetic data demonstrate that good results could be attained by applying this method to gravity data inversion compared with the least-squares regularization inversion and fully convolutional networks (FCNs).

Terrain scene clustering is a class of unsupervised methods for choosing suitable algorithms or parameters for airborne laser scanning (ALS) point cloud processing. Most existing point cloud clustering methods use hand-crafted features, such as viewpoint feature histogram (VFH), as the input of clustering algorithms. However, few studies on point cloud processing focused on terrain scene clustering via an unsupervised deep neural network. In the present study, we create a data set for terrain scene clustering in ALS point clouds. We also propose DPCC-Net, a deep point cloud clustering network via unsupervised deep learning that jointly learns the parameters of the network and the cluster task of extracted features. DPCC-Net iteratively groups the features extracted by the deep convolution neural network with the${k}$-means algorithm and uses the clustering result as the pseudo label to update the parameters of the network. We apply the proposed DPCC-Net to unsupervised training on a large terrain scene data set. The clustering result of DPCC-Net outperforms those of other typical methods.

Traditional target detection methods assume that the background spectrum is subject to the Gaussian distribution, which may only perform well under certain conditions. In addition, traditional target detection methods suffer from the problem of the unbalanced number of target and background samples. To solve these problems, this study presents a novel target detection method based on asymmetric weighted logistic metric learning (AWLML). We first construct a logistic metric-learning approach as an objective function with a positive semidefinite constraint to learn the metric matrix from a set of labeled samples. Then, an asymmetric weighted strategy is provided to emphasize the unbalance between the number of target and background samples. Finally, an accelerated proximal gradient method is applied to identify the global minimum value. Extensive experiments on three challenging hyperspectral datasets demonstrate that the proposed AWLML algorithm improves the state-of-the-art target detection performance.

Metric learning, which aims to obtain a metric matrix M such that samples from the same class are close to one another and samples of different classes are far from one another, is widely used in the field of hyperspectral dimensionality reduction (DR) and classification. Traditional metric learning is based on the Mahalanobis distance, which measures the similarity between samples via point-to-point distance, ignoring the structural features of the hyperspectral images (HSIs). To solve the above problem, we proposed clustered multiple manifold metric learning (CM3L), which obtains a manifold distance (MD), aimed at improving discrimination by introducing structural features of the HSIs and achieving good results. However, this manifold distance still has certain shortcomings in specific application situations. MD only considers the labeled data in the construction of the manifold and ignores the unlabeled data, resulting in the destruction of the manifold. Therefore, this article proposes a new spatial–spectral manifold distance (SSMD) to improve the performance of metric learning in hyperspectral DR and classification by maintaining the integrity of the constructed manifolds. The SSMD selects suitable neighboring points in the labeled and unlabeled data through the spectral–spatial information in order to participate in the construction of the manifold. Then, the distance between the manifolds is calculated to replace the traditional Mahalanobis distance. The results of seven sets of comparison experiments on three real HSI datasets demonstrate the effectiveness of SSMD in improving the classification results of HSIs.

Identification of lineament structure plays a vital role in determining the metallogenic area and distribution of the geologic structure. Edge detection methods are mostly used to recognize the lineaments and define the geologic boundaries. Cooperatively using edge detection results of the gravity, magnetic and remote sensing data to recognize lineaments would obtain more geologic information. In this paper, new edge detectors of potential field derivatives are proposed to determine the sources’ boundary, named second tilt derivative, tilt of vertical derivative, and normalized second vertical derivative, respectively. Presented approaches are characterized by producing zero amplitude over sources’ edges and equalizing anomalies from different depths. Compared with original edge detection techniques including other second derivative methods, synthetic examples reveal significant superiorities of suggested approaches in providing more accurate and sharper edges and are especially effective in distinguishing superimposed anomalies. The experiments also demonstrate that the normalization to the edge detectors will make images cleaner and geologic edges more easily captured. Applied to airborne gravimetric and magnetic data in the Pobei area (NW China), the proposed methods display more geologic details and lineaments. Canny, Sobel, and Prewitt operators are applied to extract boundaries of remote sensing image. Lineaments picked by the three different types of data are combined collectively to get a comprehensive lineaments structure interpretation.

Geothermal systems are usually characterized by distinct electrical resistivity structures due to their close relationship with the distribution of the temperature, geothermal fluids, and clay minerals. Geophysical electromagnetic (EM) methods have been routinely used for geothermal exploration. However, inversion models independently derived from different EM datasets often exhibit ambiguous resistivity features, causing difficulty in reliable interpretation. This work employs a 3-D Gauss–Newton (GN) approach to jointly invert collocated magnetotelluric (MT) and controlled-source EM (CSEM) data to better characterize subsurface resistivity structures. To effectively integrate complementary information in different datasets, we have developed a data gradient weighting adaptive joint inversion algorithm based on the norms of individual data residuals to balance contributions from each dataset during the joint inversion process. We first demonstrate the effectiveness of the developed approach on synthetic MT and CSEM data generated from a simplified geothermal model. Then, we apply the 3-D joint inversion approach to the MT and CSEM field datasets from the Yanggao region in Shanxi Province, China, to image a potential high-temperature geothermal system. Compared to results from single inversions, the joint inversion results demonstrate improved model characterization for both conductive and resistive structures, which is consistent with the results from the seismic surveys. From the joint inversion results, we can clearly delineate the important components of the geothermal system in the region, including thermal reservoirs, hydrothermal alteration layers, as well as conduction channels connecting them. Finally, we propose a conceptual model of the geothermal system in the region, which will be helpful for future drilling purposes.

The controlled-source electromagnetic (CSEM) method is a crucial tool for near-surface investigations and hydrocarbon exploration because of its economic benefits. Limited resolution is one of the inherent defects of the CSEM method, and to obtain high contrast results a multinary transform function was introduced to CSEM 3-D inversion. The multinary transform function was constituted by superposing several error functions, and the transform function transformed the model parameters from continuously distributed space into a semistep distributed multinary space. To deal with complex geometries, the edge-based finite element (FE) method with an irregular hexahedral grid was applied to the modeling and inverse problem. We used the Gauss–Newton optimization method to minimize the Tikhonov parametric function. The complex chessboard model and the marine CSEM model with complex geometry were used to validate the ability of the new method in improving the resolution of the CSEM method. By comparing the results of conventional, focusing, and multinary inversion methods, the effectiveness of the multinary inversion method in depicting sharp boundaries of different physical properties was proved. Additionally, the comparisons proved that multinary inversion method can, to some extent, overcome the insensitivity to low conductors of the CSEM method.

An efficient algorithm for the$\mathrm {L}_{ \mathrm {p}}$-norm joint inversion of gravity and magnetic data using the cross-gradient constraint is presented. The presented framework incorporates stabilizers that use$\mathrm {L}_{ \mathrm {p}}$-norms ($0\leq \mathrm {p} \leq 2$) of the model parameters, and/or the gradient of the model parameters. The formulation is developed from standard approaches for independent inversion of single data sets, and, thus, also facilitates the inclusion of necessary model and data weighting matrices, for example, depth weighting and hard constraint matrices. Using the block Toeplitz Toeplitz block structure of the underlying sensitivity matrices for gravity and magnetic models, when data are obtained on a uniform grid, the blocks for each layer of the depth are embedded in block circulant circulant block matrices. Then, all operations with these matrices are implemented efficiently using 2-D fast Fourier transforms, with a significant reduction in storage requirements. The nonlinear global objective function is minimized iteratively by imposing stationarity on the linear equation that results from applying linearization of the objective function about a starting model. To numerically solve the resulting linear system, at each iteration, the conjugate gradient algorithm is used. This is improved for large scale problems by the introduction of an algorithm in which updates for the magnetic and gravity parameter models are alternated at each iteration, further reducing total computational cost and storage requirements. Numerical results using a complicated 3-D synthetic model and real data sets obtained over the Galinge iron-ore deposit in the Qinghai province, north-west (NW) of China, demonstrate the efficiency of the presented algorithm.

The increasing complexity of seismic exploration projects and the request for higher imaging resolution have driven the geophysics community to look for a sound understanding of the subsurface formation to optimize seismic structure interpretation and reservoir characterization. Crosswell seismic survey aims at obtaining higher resolution images of the interwell regions and more accurately characterizing the reservoir dynamics. However, the presence of the intrinsic seismic attenuation of rocks as seismic waves propagate through the subsurface results in amplitude decay and velocity dispersion. This inevitably decreases the imaging resolution and the reliability of the subsequent seismic interpretation and reservoir characterization. To compensate for the attenuating effect, one may restore to attenuation compensation technique during seismic imaging. We here present the$Q$-compensated viscoelastic reverse time migration ($Q$-ERTM) based on the decoupled fractional Laplacian (DFL) viscoelastic wave equation for high-resolution crosswell imaging. We develop an explicit stabilization scheme to resolve the cumbersome numerical instability issue in$Q$-ERTM. The merits of explicit stabilization are twofold. First, it simplifies the workflows of the$Q$-ERTM by avoiding domain transforms. In addition, it provides a flexible way for stabilization parameter tuning by introducing a reference scaling factor. We follow the best practices of high-performance computing with the MPI + CUDA configuration for numerical implementation. A toy crosswell imaging example and a more realistic time-lapse crosswell seismic survey with a CO2plume injection are provided to verify the feasibility and stability of the proposed method.

Three-dimensional (3-D) inversion technique has become an important and practical approach for magnetotelluric (MT) data interpretation. In this article, we developed a 3-D parallelized MT inversion scheme using the edge-based finite element method and applied the developed method to the newly collected MT data in the Xinjiang Luntai area. The distorted hexahedral element is adopted to incorporate topography into the forward modeling and inversion for complicated scenarios. We use the Gauss–Newton optimization method to minimize the objective functional for MT inversion. The developed algorithm is parallelized using MPI over frequencies and parallel direct solvers when solving the forward and adjoint problems for each frequency. We compare the performance of the least-square QR (LSQR) factorization and preconditioned conjugate gradient (PCG) solvers for the model update within each Gauss–Newton iteration and found that the LSQR solver is more stable. The developed inversion algorithm is validated using several synthetic models. Finally, we applied the inversion algorithm to the subsurface resistivity imaging in the Luntai area. The recovered geoelectric model from full 3-D inversion fits well with the known geological and geophysical information. The recovered model shows a low resistivity layer which may be caused by the salt strata. Besides, the inversion results reveal the movement tectonic in this survey area within a depth of 9 km.

The performance of building extraction in remotely sensed images has been hugely improved with the development of convolutional neural networks and especially the semantic segmentation field. Due to the rich context of the image scene and the way of labeling (based on the pixel-level predicted probability), the segmentation masks are not always regular or close to the real building boundaries. In order to solve this problem, we propose a simple but effective deep framework based on two stages: the coarse result with an automatic semantic segmentation network and the fine result with an interactive refinement network. By using the binary mask of the initial segmentation and the interactions provided by the users, we obtain the final building extraction result through a deep interactive segmentation network. We evaluate our method on the WHU building dataset, and the results show that the method achieves better performance than the state-of-the-art methods.

Automatic extraction of buildings from high-resolution remote sensing imagery is very useful in many applications such as city management, mapping, urban planning and geographic information updating. However, due to the general texture of the building and the complexity of the image background, high-precision building segmentation from high-resolution sensing image is still a challenging task. Existing state-of-the-art frameworks use repeated pooling and step operations leading to the loss of detailed information. Thus, high-resolution representations are essential for building extraction. On this basis, our proposed network, named as HRLinkNet, maintains high-resolution representations through the whole process based on the LinkNet. We tested it on WHU Building dataset. Experimental results show that the proposed HRLinkNet is superior to the LinkNet, UNet, DLinkNet, segnet and so on.

Water-body segmentation in high-resolution satellite imagery is challenging because of the significant variations in the appearance, size, and shape of water bodies. In this letter, a novel multiscale refinement network (MSR-Net) is proposed for water-body segmentation. Similar to most learning-based methods, the MSR-Net resorts to the multiscale information for segmentation, but it improves existing networks in two ways: First, it uses the multiscale information in a new perspective. Instead of the traditional one-off manner that concatenates features and conducts segmentation on one uniform scale, the MSR-Net adopts a new multiscale refinement scheme that makes full use of the multiscale features for more accurate water-body segmentation. In addition, a novel erasing-attention module is designed for an effective feature embedding during the refinement scheme. Experiments on the Gaofen Image Data Set and the DeepGlobe Data Set demonstrate the superiority of MSR-Net when compared with the other state-of-the-art semantic segmentation methods, including U-Net, SegNet, DeepLabv3+, and ExFuse.

Semantic labeling is a fundamental task that can provide useful semantics for many other 3-D processing tasks. To tackle the challenge of airborne laser scanning (ALS) point cloud classification, current state-of-the-art methods leverage the capabilities of deep learning. However, they are limited due to the weaknesses of the isolated use of individual representations of point clouds. To address this issue, this letter presents a novel network, VPNet, which ensembles voxel and pixel representation-based networks, to predict class probabilities for each light detection and ranging (LiDAR) point. A fully connected conditional random field-based global refinement is then performed over each point in the point cloud to produce a fine-grained classification result. On the ISPRS 3-D Semantic Labeling Contest, our solution sets a new state of the art by improving the highest average F1-score and the highest average per-class accuracy from 69.3% to 73.9%, and 69.0% to 74.9%, respectively. The overall accuracy of our approach is 84.0%.

With the high spectral resolution, hyperspectral image (HSI) can provide a wealth of information for image classification. Many classification methods utilize the training samples to classify the ground materials. However, the small sample problem is still urgent to be solved when considering the cost of labeling training samples. In order to solve this problem, this letter proposes a semisupervised classification method based on the simple linear iterative cluster (SLIC) segmentation for HSI. This method improves the SLIC method to better explore the spectral characteristic of HSI. It explores the learned superpixel map and initial classification map to select the pseudo-labeled samples (PLSs), which is expected to increase the effectiveness of PLSs. The final classification map can be obtained with the integrated labeled training samples and PLSs. Experiments were carried out on three HSIs, and it was founded that the proposed method generally shows a better classification performance than the other methods.

Presently, the 3-D inversion technique has started playing a more important role in controlled-source electromagnetic (CSEM) data interpretation. With the development of hardware and computation algorithm, 3-D inversion technique has developed rapidly during the past decades. In this article, we present a newly developed 3-D parallelized inversion algorithm in the frequency domain with hexahedral discretization. Within the framework of this approach, we use the finite-element method (FEM) in the forward modeling and Gauss-Newton optimization technique in the inversion. We solve the forward modeling and adjoint problem efficiently with Math Kernel Library (MKL) Pardiso parallel direct solver. Considering the fact that the forward modeling and sensitivity calculation are frequency independent, we further parallelize the algorithm over frequency using Message Passing Interface (MPI) to speed up the modeling and inversion process. The sensitivity matrix is calculated explicitly, which enables us to estimate the optimized regularization parameter easily based on the spectral radius estimation. We proposed a new roughness operator for hexhedral discretization which works well for CSEM inversion problems. We applied the developed algorithm to several realistic CSEM models. The inversion results demonstrate the effectiveness and stability of our inversion scheme.

Forest gross primary productivity (GPP), the capacity of fixing CO2through photosynthesis, plays an important role in global changing and carbon cycle. In this study, we proposed a methodology to accurately and efficiently estimate GPP for six dominant forest types in north-east China using the remote sensing imagery driven Physiological Principles Predicting Growth model (3-PG). The GPP were accurately estimated and the results revealed that the largest GPP is obtained from the populus tremula (POTR) forest, with a range of 11.91 Mg C ha-1year-1and 40.11 Mg C ha-1year-1. While the GPP of spruce (PIAS), dahurian larch (LAGM), and fir (ABFA) are similar, with an average of 14.47 Mg C ha-1year-1, 14.85 Mg C ha-1year-1, 13.39 Mg C ha-1year-1, respectively. The GPP of korean pine (PIKO) and white birch (BEPL) are the smallest and with mean values of 7.38 Mg C ha-1year1and 3.70 Mg C ha-1year-1, respectively. In addition, the comparison with previous researches indicated that the proposed module is reasonable and credible.

The earth's natural pulse electromagnetic field (ENPEMF) signal, which is released by the instantaneous disturbance of the earth's natural changing magnetic field, contains a large amount of information about the changing geological structures and their kinetic principles. The analysis based on the time-frequency (TF) representation can provide significant interpretations for the ENPEMF signal. Wigner-Ville distribution (WVD) is one of the typical TF representations and has a high degree of TF concentration but is accompanied by severe cross-term interference. Hence, we propose a new TF representation called matching pursuit-based double WVD (MP-DWVD) to suppress cross-term interference, and synthetic signals are used to demonstrate the high TF concentration of the proposed TF analysis method. Furthermore, we apply MP-DWVD to explore the TF characteristics of the ENPEMF signal collected during the Lushan Ms7.0 earthquake. The results show that MP-DWVD can reveal accurate TF distributions of the ENPEMF signal before and after the earthquake and extract anomaly TF components, which have the potential to reveal electromagnetic anomaly information contained in the ENPEMF signal.

Target detection from hyperspectral images is an important problem. Many target detection algorithms have been proposed and have been widely used in real applications during the past decades. However, the performance of these algorithms is highly susceptible to the quality of the target spectrum. This paper proposes a multi-priori learning algorithm to learning the inherent spectral similarity and difference between multiple priori target spectra, which can alleviate the target spectral variation by boosting the priori target spectra. Experiments on two hyperspectral images illustrated the effectiveness of the proposed algorithm.

Marine controlled-source electromagnetic (CSEM) method has become a popular technique for offshore hydrocarbon exploration. It has been well recognized that marine CSEM data are strongly affected by electrical anisotropy of geologic formations in practice. Here, we present a robust and efficient finite volume algorithm for simulating marine CSEM responses in 3-D arbitrarily anisotropic formations. The algorithm is based on scattered scalar–vector potentials which improve the ill-conditioning of the resulting linear system by deflating the null space of the curl operator, and a conservative volume averaging scheme is utilized for the discretization of arbitrary electrical anisotropy. The accuracy of our algorithm is validated against quasi-analytic solutions for a layered vertical transverse isotropic reservoir model. It is then demonstrated by numerical results that the marine CSEM fields are significantly affected by the anisotropic conductivity tensor, and neglect of the full anisotropy of geologic formations may cause misleading data interpretation.

For automatic building detection in aerial images, small and dense buildings make it a very challenging task. It is because small objects lack sufficient information, and dense building distribution makes the localization of the objects confusing. High-quality building proposals can certainly promote the detection performance. The key to the problem is adopting sufficiently proper size and location of bounding boxes to use the image information for the proposal generation. Based on machine learning with a deep convolutional neural network, this letter proposes a new pipeline of building proposal generation, which is an end-to-end process during training and testing. First, the proposed pipeline attempts to find possible object center points called point proposals. Subsequently, a location refinement module and an object scoring module are applied to the boxes generated from the point proposals with a series of sizes and aspect ratios to obtain the final object proposals. This center-point-guided location refinement and multibox scoring method effectively alleviates the small and dense object problems. Experiments in INRIA Aerial Image Labeling data set demonstrate the better performance of our approach than other state-of-the-art proposal methods. In addition, we add a normal classification branch based on our generated proposals to conduct experiments on detection task. Detection result outperforms the latest detection framework R-FCN equipped with ResNet-101 7% mean average precision at 0.7.

Surface nuclear magnetic resonance (SNMR) technique has been widely applied to noninvasive groundwater exploration and aquifer quantitative characterization. Furthermore, it has the potential for studying and detecting hydrocarbon contaminants and for water resource management. To increase the tomography resolution of groundwater and expand the application range of SNMR, we first implemented its 3-D forward modeling using a vector finite-element (FE) method based on the total electric field. The key of SNMR forward modeling is the calculation of excitation magnetic field, and the total-field algorithm is an alternative scheme to numerically simulate the electromagnetic field. In this paper, we viewed the circular loop source as the combination of a certain number of horizontal electric dipoles. The unstructured tetrahedral mesh and local refinement technology were combined to precisely delineate the distribution of circular loop source, which effectively reduces the adverse effects of field source singularity. Then, the vector FE solver based on the total electric field was used to calculate the magnetic field distribution. The accuracy of the calculated results was validated by the analytical solutions for a circular loop laid on the surface of a homogeneous half-space. After calculating the excitation field, SNMR responses of three typical synthetic 3-D groundwater models were obtained with the basic signal response formula. Finally, we studied the effects of some important factors which have significant influences on SNMR signal responses and have practical importance in groundwater issues.

Target detection is playing an important role in hyperspectral image (HSI) processing. Many traditional detection methods utilize the discriminative information within all the single-band images to distinguish the target and the background. The critical challenge with these methods is simultaneously reducing spectral redundancy and preserving the discriminative information. The multitask learning (MTL) technique has the potential to solve the aforementioned challenge, since it can further explore the inherent spectral similarity between the adjacent single-band images. This letter proposes an independent encoding joint sparse representation and an MTL method. This approach has the following capabilities: 1) explores the inherent spectral similarity to construct multiple sub-HSIs in order to reduce spectral redundancy for each sub-HSI; 2) takes full advantage of the prior class label information to construct reasonable joint sparse representation and MTL models for the target and the background; 3) explores the great difference between the target dictionary and background dictionary with different regularization strategies in order to better encode the task relatedness for two joint sparse representation and MTL models; and 4) makes the detection decision by comparing the reconstruction residuals under different prior class labels. Experiments on two HSIs illustrated the effectiveness of the proposed method.

Automatic cloud detection from satellite imagery is a necessary preprocessing step in remote sensing. Given that humans can easily “see” clouds in an image because of salient region features, we adopt a visual attention technique in computer vision to automatically identify images with a significant cloud cover. The proposed method generates a rough cloud mask by using a top-down visual saliency model to qualitatively distinguish cloud images from noncloud images. First, an image is downsized for rapid processing. Some basic saliency maps of clouds are then generated by multilevel segmentation, the computation of cloud visual saliency features, and feature classification. Thereafter, we fuse the basic saliency maps by using a most-votes-win strategy to generate the cloud mask. With the cloud mask, a threshold is used to classify the images as cloud or noncloud images. A total of 200 RapidEye images are tested by using the algorithm. Of the cloud images, 92% are correctly identified. The average processing time is 1.8 s per image.

Automatic extraction of roads from images of complex urban areas is a very difficult task due to the occlusions and shadows of contextual objects, and complicated road structures. As light detection and ranging (LiDAR) data explicitly contain direct 3-D information of the urban scene and are less affected by occlusions and shadows, they are a good data source for road detection. This paper proposes to use multiple features to detect road centerlines from the remaining ground points after filtering. The main idea of our method is to effectively detect smooth geometric primitives of potential road centerlines and to separate the connected nonroad features (parking lots and bare grounds) from the roads. The method consists of three major steps, i.e., spatial clustering based on multiple features using an adaptive mean shift to detect the center points of roads, stick tensor voting to enhance the salient linear features, and a weighted Hough transform to extract the arc primitives of the road centerlines. In short, we denote our method as Mean shift, Tensor voting, Hough transform (MTH). We evaluated the method using the Vaihingen and Toronto data sets from the International Society for Photogrammetry and Remote Sensing Test Project on Urban Classification and 3-D Building Reconstruction. The completeness of the extracted road network on the Vaihingen data and the Toronto data are 81.7% and 72.3%, respectively, and the correctness are 88.4% and 89.2%, respectively, yielding the best performance compared with template matching and phase-coded disk methods.

The fast filtering of massive point cloud data from light detection and ranging (LiDAR) systems is important for many applications, such as the automatic extraction of digital elevation models in urban areas. We propose a simple scan-line-based algorithm that detects local lowest points first and treats them as the seeds to grow into ground segments by using slope and elevation. The scan line segmentation algorithm can be naturally accelerated by parallel computing due to the independent processing of each line. Furthermore, modern graphics processing units (GPUs) can be used to speed up the parallel process significantly. Using a strip of a LiDAR point cloud, with up to 48 million points, we test the algorithm in terms of both error rate and time performance. The tests show that the method can produce satisfactory results in less than 0.6 s of processing time using the GPU acceleration.

Automatic detection of regions of salient texture and objects is useful for analysis of remotely sensed imagery, such as for land cover classification, object detection, and change detection. Intuitively, the local edges on an image indicate spectral discontinuity and the existence of structure texture or objects. This letter explores a simple method for measuring the saliency of texture and objects based on the edge density and spatial evenness of the edge distribution in the local window of each pixel. This method generates a saliency map by computing the saliency index of each pixel. By segmenting the saliency map, the salient structure texture regions and the locations of objects can be extracted. The algorithm requires only the window size as the input parameter and is relatively simple to implement. Experiments using high-resolution images show its effectiveness and accuracy in the detection of salient structure texture regions, such as crops and residential areas, and man-made objects, such as airplanes, cars, etc.