IEEE Transactions on Geoscience and Remote Sensing

This article proposes an innovative scheme for recovering sparse reflectivity series from uniformly quantized seismic signals. In this scheme, the statistically less affected impulses by the quantization error are assigned higher weights than the ones with a larger error. First, the orthogonal matching pursuit (OMP) algorithm is applied on a quantized seismic trace to obtain a set of conservative estimates of the reflectivity impulses. Second, the quantization error is formulated as a systematic uncertainty within a neighborhood of the obtained conservative estimates from the OMP. Finally, a robust worst case (RWC) deconvolution method is developed to recover an improved estimate of the reflectivity impulses. The proposed scheme significantly increases the robustness and enhances the recovered reflectivity impulses obtained by the OMP algorithm. This is substantiated by experiments on both synthetic and real seismic data. Specifically, the falsely or overly estimated impulses are significantly suppressed, and the robustness to the change of the quantization interval is enhanced.

The angular and spectral kernel-driven (ASK) model distinguishes soil and vegetation spectral features by the component spectra and is a promising model which combines multisensor data for inversion. However, its global application is limited by the component spectra. This article proposes parameterization of the ASK component spectra of soil and leaf from global spectra libraries as ANGERS, GOSPEL, LOPEX, and USGS. A statistical ratio (y) of various leaf to soil spectra is used to capture their spectral differences and variations, with mean (m) + u (0, ±0.5, ±1) standard deviations (σ) [i.e., y (m + uσ)]. Optimization inversion is applied to determine the ratio candidates y(m + uσ), allowing more tolerance for spectral uncertainty, which releases the semiempirical nature of the kernel-driven model. Simulation data analysis proves its feasibility and good capture of vegetation-soil spectral differences. The model's bidirectional reflectance factor (BRF) fitting error [root-mean-square error (RMSE)] of 0.0245 is slightly larger than the true component spectra of 0.0178, and albedo RMSE is 0.0116 in Black Sky Albedo and 0.0182 in White Sky Albedo. The result also shows its good robustness to the noises, where the====level up to 20% noise conducts a 0.0277 error in BRF fitting and an ignorable influence in albedo. The synergistic-retrieved albedo from multisensor satellite data consists of in situ measurements with an RMSE of 0.0171, compared to 0.0131 from true component spectra retrievals. The new parameterization sacrifices some accuracy, but it is simple and operational for global retrieval with a satisfactory precision.

Existing methods of the small target detection from infrared videos are not effective with the complex background. It is mainly caused by: 1) the interference of strong edges and the similarity with other nontarget objects and 2) the lack of the context information of both the background and the target in a spatio-temporal domain. By considering these two points, we propose to slide a window in a single frame and form a spatio-temporal cube with the current frame patch and other frame patches in the spatio-temporal domain. Then, we establish a spatio-temporal tensor model based on these patches. According to the sparse prior of the target and the local correlation of the background, the separation of the target and the background can be cast as a low rank and sparse tensor decomposition problem. The target is obtained from the sparse tensor by the tensor decomposition. The experiments show that our method gains better detection performance in infrared videos with the complex background by making full use of the spatio-temporal context information.

In this article, an automatic and forward method is realized to establish attributed scattering center models directly from the computer-aided design (CAD) model of the complex target. The main steps include the preprocessing of the CAD model, the separation of scattering sources, the selection of strong scattering sources, and the automatic determination of model parameters. With the proposed method, the scattering sources, scattering mechanisms, and model parameters of the scattering centers can be identified and derived, such that the complicated manual intervention is completely avoided. Moreover, the method establishes the distributed scattering center models formed by curved surfaces with large curvature radii. Therefore, the formation mechanism of the distributed scattering center is extended from typical scattering structures to a general case. Thus, the model of the attributed scattering center is extended and can be applied to describe the scattering from the real structures of the complex target. The geometric shape of the scattering source is distinguished based on the principal curvature radii, which are calculated by differential geometry theory. Thus, the frequency dependence parameter is obtained according to its corresponding relationship with the geometric shape. In addition, based on the automatic method, a technology is studied to diagnose and correct the scattering center models of a target whose CAD model is unknown or partially known. Finally, parametric models of several targets in the Moving and Stationary Target Acquisition Recognition (MSTAR) program are established, and then compared with the measured data. The results validate the effectiveness of the proposed method.

Primary productivity (PP) has been recently investigated using remote sensing-based models over quite limited geographical areas of the Red Sea. This work sheds light on how phytoplankton and primary production would react to the effects of global warming in the extreme environment of the Red Sea and, hence, illuminates how similar regions may behave in the context of climate variability. study focuses on using satellite observations to conduct an intercomparison of three net primary production (NPP) models-the vertically generalized production model (VGPM), the Eppley-VGPM, and the carbon-based production model (CbPM)-produced over the Red Sea domain for the 1998-2018 time period. A detailed investigation is conducted using multilinear regression analysis, multivariate visualization, and moving averages correlative analysis to uncover the models' responses to various climate factors. Here, we use the models' eight-day composite and monthly averages compared with satellite-based variables, including chlorophyll-a (Chla), mixed layer depth (MLD), and sea-surface temperature (SST). Seasonal anomalies of NPP are analyzed against different climate indices, namely, the North Pacific Gyre Oscillation (NPGO), the multivariate ENSO Index (MEI), the Pacific Decadal Oscillation (PDO), the North Atlantic Oscillation (NAO), and the Dipole Mode Index (DMI). In our study, only the CbPM showed significant correlations with NPGO, MEI, and PDO, with disagreements relative to the other two NPP models. This can be attributed to the models' connection to oceanographic and atmospheric parameters, as well as the trends in the southern Red Sea, thus calling for further validation efforts.

The spatiotemporal coverage of a Moon-based synthetic aperture radar (SAR) is analyzed based on the imaging geometry, upon which the spatial coverage and image formulation rely. The distance from the Earth to the Moon-based SAR and bounds of the grazing and azimuthal angles jointly determine the coverage area on the Earth’s surface. Meanwhile, the ground coverage of the Moon-based SAR is determined by the bounds of the grazing and azimuthal angles and geographic coordinates of the nadir point at a specified time. Moreover, the temporal variation in the spatial coverage is pertinent to the temporally varying nadir point of the Moon-based SAR on the Earth’s surface. Furthermore, numerical simulations using the lunar ephemeris data are carried out to complement the analysis and to illustrate the spatiotemporal coverage. Finally, a guideline for the optimal site selection of a Moon-based SAR is proposed. In conclusion, a Moon-based SAR has the potential to perform long-term, continuous Earth observations on a global scale to enhance our capability to understand the planet.

In the case of sparse aperture, the coherence between pulses of radar echo is destroyed, which challenges inverse synthetic aperture radar (ISAR) autofocusing and imaging. Mathematically, reconstructing the ISAR image from the sparse aperture radar echo is a linear underdetermined inverse problem, which, by nature, can be solved by the fast developed compressive sensing (CS) or sparse signal recovery theory. However, the CS-based sparse aperture ISAR imaging algorithms are generally computationally heavy, which becomes the bottleneck of preventing their applications to the real-time ISAR imaging system. In this article, we propose a novel and computationally efficient ISAR autofocusing and imaging algorithm for sparse aperture. We first consider a generalized CS model for ISAR imaging and autofocusing with sparse and entropy-minimization regularizations, and then utilize the alternating direction method of multipliers (ADMM) algorithm to optimize the model. To improve computational efficiency, the matrix inversion is translated to an elementwise division with the usage of a partial Fourier dictionary, and the 2-D ISAR image is updated as a whole instead of range cellwise. To achieve autofocusing for sparse aperture, the phase error is estimated by minimizing the entropy of the ISAR image reconstructed in each iterative loop. Experiments based on both simulated and measured data validate that the proposed algorithm can achieve well-focused ISAR images within a few seconds, which is ten times faster than the reported sparse aperture ISAR imaging algorithms.

Spectral unmixing is an important task in hyperspectral image (HSI) analysis and processing. Sparse representation has become a promising semisupervised method for remotely sensed hyperspectral unmixing and incorporating the spectral or spatial information to improve the spectral unmixing results under a weighted sparse unmixing framework is a recent trend. While most methods focus on analyzing HSI by exploring the spatial information, it is known that hyperspectral data are characterized by its large contiguous set of wavelengths. This information can be naturally used to improve the representation of pixels in HSI. In order to take the advantage of the hyper spectral information as well as the spatial information for hyperspectral unmixing, in this article, we explore and introduce a multiview data processing approach through spectral partitioning to benefit from the abundant spectral information in HSI. Some important findings on the application of multiview data set in sparse unmixing are discussed. Meanwhile, we develop a new spectral-spatial-weighted multiview collaborative sparse unmixing (MCSU) model to tackle such a multiview data set. The MCSU uses a weighted sparse regularizer, which includes both multiview spectral and spatial weighting factors to further impose sparsity on the fractional abundances. The weights are adaptively updated associated with the abundances, and the proposed MCSU can be solved by the alternating direction method of multipliers efficiently. The experimental results on both the simulated and real hyperspectral data sets demonstrate the effectiveness of the proposed MCSU, which can significantly improve the abundance estimation results.

High/very-high-resolution (HR/VHR) multitemporal images are important in remote sensing to monitor the dynamics of the Earth’s surface. Unsupervised object-based image analysis provides an effective solution to analyze such images. Image semantic segmentation assigns pixel labels from meaningful object groups and has been extensively studied in the context of single-image analysis, however not explored for multitemporal one. In this article, we propose to extend supervised semantic segmentation to the unsupervised joint semantic segmentation of multitemporal images. We propose a novel method that processes multitemporal images by separately feeding to a deep network comprising of trainable convolutional layers. The training process does not involve any external label, and segmentation labels are obtained from the argmax classification of the final layer. A novel loss function is used to detect object segments from individual images as well as establish a correspondence between distinct multitemporal segments. Multitemporal semantic labels and weights of the trainable layers are jointly optimized in iterations. We tested the method on three different HR/VHR data sets from Munich, Paris, and Trento, which shows the method to be effective. We further extended the proposed joint segmentation method for change detection (CD) and tested on a VHR multisensor data set from Trento.

The first rotating fan beam scatterometer onboard China-France Oceanography Satellite (CFOSAT) was successfully launched on October 29, 2018. CFOSAT SCATterometer (CSCAT) is dedicated to the monitoring of sea surface wind vectors but also provides valuable data for the applications over land and Polar Regions. This article provides an overview of the relevant procedures of CSCAT data processing, including onboard signal processing and operational ground processing. Then a post-launch analysis is carried out to evaluate the first results of CSCAT L1 and L2 products. It shows that the CSCAT instrument is generally stable in terms of noise measurements and internal calibration, unless there is any important change in the system configuration. Specifically, the CSCAT backscatter (σθ) precision and wind quality are studied using a set of collocated ancillary data. The σθ precision degrades as wind speed decreases, and it is relatively low at high incidence angles (e.g., θ >46°). In particular, backscatter estimation of the horizontally polarized beam should be further improved by correcting the noise subtraction factor. The retrieved CSCAT winds are in good agreement with the European Centre for Medium Range Weather Forecasts (ECMWF) winds, the Advanced Scatterometer (ASCAT) winds, as well as the buoy winds. However, due to unresolved calibration and interbeam consistency problems, the wind quality degrades remarkably for the out-swath and the nadir-region wind vector cells, implying that the σθ calibration should be improved in the future updates.

Traditionally, clean reference images are needed to train the networks when applying the deep learning techniques to tackle image denoising tasks. However, this idea is impracticable for the task of synthetic aperture radar (SAR) image despeckling, since no real-world speckle-free SAR data exist. To address this issue, this article presents a noisy reference-based SAR deep learning filter, by using complementary images of the same area at different times as the training references. In the proposed method, to better exploit the information of the images, parameter-sharing convolutional neural networks are employed. Furthermore, to mitigate the training errors caused by the land-cover changes between different times, the similarity of each pixel pair between the different images is utilized to optimize the training process. The outstanding despeckling performance of the proposed method was confirmed by the experiments conducted on several multitemporal data sets, when compared with some of the state-of-the-art SAR despeckling techniques. In addition, the proposed method shows a pleasing generalization ability on single-temporal data sets, even though the networks are trained using finite input-reference image pairs at a different imaging area.

Urban road extraction has wide applications in public transportation systems and unmanned vehicle navigation. The high-resolution remote sensing images contain background clutter and the roads have large appearance differences and complex connectivities, which makes it a very challenging task for road extraction. In this article, we propose a novel end-to-end deep learning model for road area extraction from remote sensing images. Road features are learned from three levels, which can remove the distraction of the background and enhance feature representation. A direction-aware attention block is introduced to the deep learning model for keeping road topologies. We compare our method on public remote sensing data sets with other related methods. The experimental results show the superiority of our method in terms of road extraction and connectivity preservation.

Imaging hydraulic fractures is of paramount importance to subsurface resource extraction, geologic storage, and hazardous waste disposal. The use of electrically conductive proppants and current energized steel casing provides a promising approach to monitor the distribution of fractures. In this article, a borehole-to-surface system is employed to energize the steel casing and measure electric and magnetic fields on the ground. A convolutional neural network (CNN) is then trained to learn the relationship between the measured field pattern and the parameterized fracture, namely, the lateral extent and direction. To accelerate the generation of training data with limited accuracy loss, an approximate hollow casing is modeled by the impedance transition boundary condition with a tenfold magnified radius and reduced conductivity. Two training strategies are then presented with a grid search of the network’s hyperparameters. The well-trained CNN shows good generalization to unseen fracture conductivity, the true casing model, as well as white Gaussian noise. Finally, we apply the CNN to image irregular fractures and obtain reliable results even under strong noise, indicating a promising imaging technique for more complicated fractures.

Hyperspectral image (HSI) unmixing is an important issue of research due to its effect on the subsequent processing of HSIs. Recently, the sparse regression method with spatial information has been successfully applied in hyperspectral unmixing (HU). However, most sparse regression methods ignore the difference in spatial structure handling with only one sparse constraint. In fact, the pixels in detail regions are more likely to be severely mixed with more endmembers participated, and the sparsity degree of its corresponding abundances is relatively low. Considering the sparsity difference of abundances, a sketch-based region adaptive sparse unmixing applied to HSI is proposed in this article. Inspired by the vision computing theory, we use the region generation algorithm based on a sketch map to differentiate the homogeneous regions and detail regions. Then, the abundances of these two kind regions in HSIs are separately constrained by sparse regularizers of ${L}_{1/2}$ and ${L}_{1}$ with a proposed manifold constraint. Our method not only makes full use of the spatial information in HSIs but also exploits the latent structure of data. The encouraging experimental results on three data sets validate the effectiveness of our method for HU.

Ground-penetrating radar inspection of vertical structures, such as columns or pillars, is relevant in several applicative contexts. Unlike conventional subsurface prospecting, where the medium is accessible only from one side, the columns can be probed from various sides with measurement domains possibly encircling the structure. This makes it possible to retrieve more information about the scene, thanks to an increased view and data collection diversity. This article proposes an imaging approach for structures probed all around via vertical scans. The approach faces the imaging as a full 3-D electromagnetic inverse scattering problem and accounts for the vectorial nature of the scattering phenomenon. Moreover, the imaging approach is based on an approximate model of scattering and the inversion is regularized by means of the truncated singular value decomposition to produce stable and accurate results. The reconstruction capabilities of the proposed imaging approach are evaluated in terms of the achievable spatial resolution. To this end, a numerical analysis exploiting synthetic data allows investigating how the imaging quality depends on the number of vertical scans. Reconstruction results referred to data gathered in controlled conditions provide an experimental assessment of the achievable imaging capabilities.

Incoherent noise is one of the most common noise widely distributed in seismic data. To improve the interpretation accuracy of the underground structure, incoherent noise needs to be adequately suppressed before the final imaging. We propose a novel method for suppressing seismic incoherent noise based on the robust low-rank approximation. After the Hankelization, seismic data will show strong low-rank features. Our goal is to obtain the stable and accurate low-rank approximation of the Hankel matrix and then reconstruct the denoised data. We construct a mixed model of the nuclear norm and the $l_{1}$ norm to express the low-rank approximation of the Hankel matrix constructed in the frequency domain. Essentially, the adopted model is an optimization for the subspace similar to the online subspace tracking method, thus avoiding the time-consuming singular value decomposition (SVD). We introduce the orthonormal subspace learning to convert the nuclear norm to the $l_{1}$ norm to optimize the orthonormal subspace and the corresponding coefficient. Finally, two optimization strategies—the alternating direction method and the block coordinate descent method—are applied to obtain the optimized orthonormal subspace and the corresponding coefficient for representing the low-rank approximation of the Hankel matrix. We perform incoherent noise attenuation tests on synthetic and real seismic data. Compared with other denoising methods, the proposed method produces small signal errors while effectively suppressing the seismic incoherent noise and has a high computational efficiency.

With a small number of labeled samples for training, it can save considerable manpower and material resources, especially when the amount of high spatial resolution remote sensing images (HSR-RSIs) increases considerably. However, many deep models face the problem of overfitting when using a small number of labeled samples. This might degrade HSR-RSI retrieval accuracy. Aiming at obtaining more accurate HSR-RSI retrieval performance with small training samples, we develop a deep metric learning approach with generative adversarial network regularization (DML-GANR) for HSR-RSI retrieval. The DML-GANR starts from a high-level feature extraction (HFE) to extract high-level features, which includes convolutional layers and fully connected (FC) layers. Each of the FC layers is constructed by deep metric learning (DML) to maximize the interclass variations and minimize the intraclass variations. The generative adversarial network (GAN) is adopted to mitigate the overfitting problem and validate the qualities of extracted high-level features. DML-GANR is optimized through a customized approach, and the optimal parameters are obtained. The experimental results on the three data sets demonstrate the superior performance of DML-GANR over state-of-the-art techniques in HSR-RSI retrieval.

With the development of convolutional neural networks (CNNs), the semantic understanding of remote sensing (RS) scenes has been significantly improved based on their prominent feature encoding capabilities. While many existing deep-learning models focus on designing different architectures, only a few works in the RS field have focused on investigating the performance of the learned feature embeddings and the associated metric space. In particular, two main loss functions have been exploited: the contrastive and the triplet loss. However, the straightforward application of these techniques to RS images may not be optimal in order to capture their neighborhood structures in the metric space due to the insufficient sampling of image pairs or triplets during the training stage and to the inherent semantic complexity of remotely sensed data. To solve these problems, we propose a new deep metric learning approach, which overcomes the limitation on the class discrimination by means of two different components: 1) scalable neighborhood component analysis (SNCA) that aims at discovering the neighborhood structure in the metric space and 2) the cross-entropy loss that aims at preserving the class discrimination capability based on the learned class prototypes. Moreover, in order to preserve feature consistency among all the minibatches during training, a novel optimization mechanism based on momentum update is introduced for minimizing the proposed loss. An extensive experimental comparison (using several state-of-the-art models and two different benchmark data sets) has been conducted to validate the effectiveness of the proposed method from different perspectives, including: 1) classification; 2) clustering; and 3) image retrieval. The related codes of this article will be made publicly available for reproducible research by the community.

Accurate and up-to-date road maps are of great importance in a wide range of applications. Unfortunately, automatic road extraction from high-resolution remote sensing images remains challenging due to the occlusion of trees and buildings, discriminability of roads, and complex backgrounds. To address these problems, especially road connectivity and completeness, in this article, we introduce a novel deep learning-based multistage framework to accurately extract the road surface and road centerline simultaneously. Our framework consists of three steps: boosting segmentation, multiple starting points tracing, and fusion. The initial road surface segmentation is achieved with a fully convolutional network (FCN), after which another lighter FCN is applied several times to boost the accuracy and connectivity of the initial segmentation. In the multiple starting points tracing step, the starting points are automatically generated by extracting the road intersections of the segmentation results, which then are utilized to track consecutive and complete road networks through an iterative search strategy embedded in a convolutional neural network (CNN). The fusion step aggregates the semantic and topological information of road networks by combining the segmentation and tracing results to produce the final and refined road segmentation and centerline maps. We evaluated our method utilizing three data sets covering various road situations in more than 40 cities around the world. The results demonstrate the superior performance of our proposed framework. Specifically, our method’s performance exceeded the other methods by 7% and 40% for the connectivity indicator for road surface segmentation and for the completeness indicator for centerline extraction, respectively.

We present a machine learning approach to classify the phases of surface wave dispersion curves. Standard frequency-time analysis (FTAN) analysis of seismograms observed on an array of receivers is converted into an image, of which each pixel is classified as fundamental mode, first overtone, or noise. We use a convolutional neural network (U-Net) architecture with a supervised learning objective and incorporate transfer learning. The training is initially performed with synthetic data to learn coarse structure, followed by fine-tuning of the network using approximately 10% of the real data based on human classification. The results show that the machine classification is nearly identical to the human picked phases. Expanding the method to process multiple images at once did not improve the performance. The developed technique will facilitate the automated processing of large dispersion curve data sets.