Azimuth ambiguity in azimuth multichannel system (AMCS) results from nonuniform sampling and degrades the synthetic aperture radar (SAR) imaging quality. In addition, the uniform pulse transmission pattern adopted in most AMCS is difficult to balance the azimuth resolution and range swath width performance. Furthermore, the design of pulse transmission pattern and reconstruction algorithm are often considered as two independent problems, which is difficult to achieve the optimal high-resolution and wide-swath (HRWS) imaging performance. In this article, we propose a joint optimization network to learn the nonuniform pulse transmission pattern and train reconstruction network parameters simultaneously, dubbed PTIR-Net. Specifically, we develope a nonuniform multichannel imaging model, consisting of the nonuniform approximate measurement operators with continuously defined pulse transmission time. To facilitate a more stable and generalized network, we introduce the implicit regularization learned by a model-based deep learning image reconstruction scheme to process the nonuniform sampling data. Extensive experiments on Gaofen-3 dataset show that the proposed network achieves better reconstruction performance in terms of both quantitative metrics and visual quality.

This letter proposes a novel self-calibration method for channel phase error (CPE) in antenna arrays by leveraging the geometric properties of non-collinear configurations. The CPE can be decomposed into linear and orthogonal components relative to the baseline length vector, which cause direction-of-arrival (DOA) bias and manifold distortion, respectively. However, for self-calibration methods, only the manifold distortion can be perceived and corrected. In collinear arrays, the DOA bias is independent of manifold distortion, so the linear component of CPE can't be self-calibrated. Fortunately, in non-collinear arrays, we can couple the DOA bias into the manifold distortion by reformulating the slant range formula with the Fresnel and virtual collinear array (VCA) approximations, making it possible to estimate the entire CPE without external references. We then propose an iterative least-squares algorithm that corrects for CPE using multiple snapshots collected from a non-collinear array. Simulations under various SNRs, distances, linear components of CPE, degrees of non-collinearity, and fields of view demonstrate the effectiveness and robustness of the proposed method.

The combination of microwave photonic (MWP) technology and synthetic aperture radar (SAR) facilitates the generation, transmission, and reception of ultra-wideband (UWB) signals, thereby enabling the production of centimeter-resolution SAR images. Recently, an experimental MWP SAR system with a bandwidth of 14 GHz has been constructed at Aerospace Information Research Institute, Chinese Academy of Sciences, which has been tested by outfield SAR experiments. In the context of this system, we analyzed the manifestation of systematical errors in the coarse-focused image and found that deviations occur in both the envelope and the phase of the range cell migration (RCM) curves, which can be considered as inherent slant range error and center frequency error. In addition, the inclusion of a radio frequency power amplifier (PA) in this system can introduce frequency-dependent error. In this article, we analyze the sources and establish signal models for the above three types of systematical errors individually and propose a systematical error estimation method by exploiting the trihedral reflectors (TRCs) within the coarse-focused image. Utilizing the error estimation results and the measured frequency response of the PA, a post-processing compensation method is also proposed. The estimation and compensation method is applied to the processing of the vehicle-borne data acquired by this experimental MWP SAR system, yielding superior performance and a fine-focused SAR image with a resolution of approximately 0.013 m in the range dimension.

To address the critical challenge of simultaneous interpretation for hybrid scattering targets in complex urban environments, this paper proposes a novel polarimetric information-based three-dimensional (3D) imaging framework. By integrating polarimetric decomposition with morphological operations, this framework achieves preliminary classification between dense manmade regions and natural terrains. For the layover of manmade targets, we propose a polarization-based joint sparse method. It can pre-judge the dominant scatterer within pixel blocks and selectively extract optimal polarization channels to facilitate layover separation by leveraging differences across various polarization channels. For distributed natural terrains, we employ a spectral analysis method based on polarimetric covariance matrix (CM) estimation. It extracts non-local homogeneous pixel blocks within the polarization-Euclidean space through the synergistic integration of scattering mechanism measurements and statistical elevation priors, thereby enabling CM and elevation estimation. Subsequently, from both simulated and real-data experiments, the proposed framework has been validated effective in enhancing reconstruction accuracy and completeness through optimal utilization of polarization information.

Monocular depth estimation aims to predict the corresponding scene depth map to an input image, which has wide application prospects in various fields, such as robot navigation, autonomous driving, and augmented reality. Due to the advantage that only images rather than ground truth depth maps are required for model training, self-supervised monocular depth estimation methods have received more and more attention in recent years. Although numerous self-supervised monocular depth estimation methods were proposed, there has been no a comprehensive survey on them yet. Addressing this issue, we review recent developments in the community of self-supervised monocular depth estimation in this article. First, 89 existing works in the literature are categorized and reviewed. Then, we introduce the public datasets and evaluation metrics used in monocular depth estimation. Next, the performances of some state-of-the-art methods are compared and analyzed. Finally, we summarize several open problems and possible future developments in this community.

The combination of microwave photonic (MWP) technology and synthetic aperture radar (SAR) can realize ultra-wideband transmitted signals and thus provides ultra-high-resolution. Recently, an experimental MWP SAR system has been implemented at Aerospace Information Research Institute, Chinese Academy of Sciences. It can generate transmitting signals up to 14 GHz bandwidth and adopts dechirp-on-receive method at the receiving end. However, the power amplifier (PA) in front of the transmitting antenna induces an additional frequency-dependent error that needs to be compensated during signal processing. In this paper, we compensate for this specific systematical error utilizing the residual video phase (RVP) removal technique. Some simulated and experimental data processing results are presented to validate the effectiveness of the method.

The current SAR ship detection datasets follow the traditional paradigm in computer vision field, that is, the image data and the corresponding annotations are provided. However, due to the special imaging mechanism of SAR, the amplitude or intensity data is limited and cannot offer sufficient electromagnetic information of target. Although the single-look complex SAR data is informative, it requires a large number of storage space and is difficult to interpret visually. To this end, a new benchmark for SAR ship detection is proposed in this paper, namely PhySAR-Det. It is constructed based on different levels of SAR products of Gaofen-3 satellite, including L1A complex-valued image and the geo-coded L2 product. We propose another microwave visual characteristic (MVC) product derived from L1A data and projected to L2 coordinates to characterize the scattering mechanisms. The L2 images with high visual quality attached with the corresponding MVC products form the proposed PhySAR-Det dataset. The experiments show that the MVC product improves the performance compared with only image data available. It will be available at https://github.com/XAI4SAR/PhySARDet.

Synthetic Aperture Radar Tomography (TomoSAR) thre-dimensional imaging technology is built upon the foundation of two-dimensional SAR imaging, utilizing multiple observations in the elevation direction to construct the synthetic aperture for the elevation imaging capability. The classic TomoSAR imaging algorithm typically refers to the third dimension as elevation, and research over the years has been based on this model. The process of three-dimensional imaging involves reconstructing the spatial distribution of scattering characteristics along the elevation direction. According to this model, for the same range-azimuth cell, discrimination of overlapping scatterers can be achieved based on different levels of sparsity. However, studies have found significant limitations of this model for urban buildings. Firstly, there is scatterer diffusion along the elevation direction, particularly at the junctions of building surfaces, such as the intersections between the facade and ground. Secondly, the maximum unambiguous range of the reconstruction model based on elevation is limited by the model. To address this issue, considering the characteristics of urban buildings, this paper proposes a novel imaging model based on the unique angle property. Specifically, the scatterer parameters to be estimated are transformed from the elevation coordinates to the off-nadir angle. Moreover, based on the non-penetrating property of electromagnetic waves for buildings, it is assumed that there is only one scatterer for each off-nadir angle, which is consistent with physical reality. Experiments were conducted based on measured data from two sites, and the results confirmed that the proposed model can effectively suppress clutter caused by multiple scattering, significantly improving the three-dimensional imaging quality.

SAR Tomography can provide three-dimensional (3D) image of the observed scenes, becoming an important technology for urban mapping, target detecting and many other fields. When imaging urban scenes, multipath signals are always considered as noise in the previous research on 3D SAR imaging. However, the signals experienced multiple bounces hold massive information about the target scenario. In this paper, we aim to turn "waste" into treasure by digging for the hidden information in multipath. Based on the TomoSAR mechanism, we complete the restoration of the 3D imaging result. Firstly, we introduce the TomoSAR imaging model under the low-altitude case, which is more suitable for airborne or UAV- borne platform in urban area. We analyze the multipath model and propose the multipath detection and restoration methods. The virtual targets caused by multipath can be restored to the correct positions in the image by relocating. To verify the validity of our method, we conducted a series of real-data experiments, the results show that our method makes full use of multipath signals to retrieve non-line-of-sight targets missing in traditional direct-path 3D imaging.

The rapid development in SAR three-dimensional imaging technology and the feasibility of employing lightweight and convenient platforms have increased the attention of low-altitude UAV-borne TomoSAR. However, the building obstructions in urban area and the low altitude cause multipath phenomenon, which is challenging in imaging interpretation and accurate reconstruction. While previous research is mainly about the removal of multipath effects, we focus on exploiting multipath information within the framework of low-altitude TomoSAR. This paper aims to explore effective strategies for determining the viability of multipath exploitation in urban canyons. Firstly, we introduce the low-altitude TomoSAR imaging model under the spherical wavefront model. Through a comprehensive analysis of the range profile of the multipath mechanism, our objective is to provide detection and mechanism determination before relocating or restoring the multipath for TomoSAR. This algorithm was applied in real-data experiments to investigate the multipath mechanisms of different cases, validating practical value.

Synthetic Aperture Radar (SAR) is extensively employed in earth remote sensing, including both civilian and military sectors. Currently, establishing the correspondence between strong scattering regions in SAR images and the 3D geometry of the target has become a research hotspot. Traditional methods based on scattering center modeling overly rely on prior knowledge and manual annotation. However, neural network-based approaches often lack the constraints of SAR imaging principles when performing feature extraction. In this paper, we propose a method for analyzing strong scattering in SAR images based on a differentiable simulator to visualize the correspondence between strong scattering in SAR images and the 3D geometry of the target. Specifically, the differentiable simulator accumulates the scattering intensity in the scattering simulation stage to generate simulated SAR images at multi-pose. By computing the loss with the SAR image in the dataset (ground truth), the differentiable simulator can provide a direct mapping from the strong scattering in SAR images to the target's 3D structures. The effectiveness of the method is validated on a dataset of simulated SAR images based on the MSTAR. From the result we obtained, the correspondence between the 3D structures of the target model and the strong scattering in the image can be intuitively and clearly determined.

Compared with optical images, SAR images exhibit significant differences in the characteristics of scenes or targets due to differences in imaging mechanisms. In particular, the layover phenomenon in SAR images leads to poor interpretability, which hinders the application of SAR products. In this paper, a three-dimensional fusion visualization method combining SAR image data and optical oblique photogrammetry data is proposed. Leveraging the three-dimensional imaging capability of microwave vision SAR, the three-dimensional characteristics of the targets are acquired by a small number of coherent observations, and fused with oblique photogrammetry models for display, thus the multidimensional target characteristics of SAR can be effectively organized and expressed. The method has strong novelty and promotes the interpretation of SAR images and product applications.

The fully polarimetric array interferometric SAR (Pol-array-InSAR) technology contributes to the three-dimensional (3D) reconstruction and characterization of complex urban areas. However, current processing methods have suffered from a substantial loss of polarization information. In response to this issue, we propose a 3D imaging methodology. This approach first projects the original SAR images into an optimal polarization space with maximum polarization energy. Then, the resulting images are used to retrieve the elevation of targets by utilizing the sparse Bayesian learning (SBL) algorithm. Supported by a four-channel Pol-array-InSAR dataset acquired by our unmanned aerial vehicle (UAV) borne system, we validate the feasibility of this methodology and compare its performance with traditional compressive sensing (CS) algorithms.

Synthetic Aperture Radar (SAR) 3D point cloud reconstruction improves target identification by mitigating issues like overlay masking and shadows found in 2D SAR image projections. Nevertheless, challenges arise from limited SAR data availability and complexities in deciphering and labeling, hindering research in SAR 3D reconstruction point cloud semantic segmentation. To address these hurdles, we propose an alternative training approach—shifting from LiDAR to SAR point clouds. Leveraging benchmark datasets in the LiDAR domain, we advocate using LiDAR point clouds for training to counter the scarcity of SAR training sets. However, applying a segmentation model across different domains leads to a performance decline, particularly in cross-modal SAR-reconstructed point clouds, attributed to distinct roughness introduced by outliers. This paper introduces a 3D semantic segmentation framework based on unsupervised domain adaptation (UDA) for cross-modal learning from LiDAR to SAR. Additionally, we present a simple yet effective geometric transformation data augmentation technique to handle highly imbalanced data distribution. Experimental results confirm the feasibility and effectiveness of our proposed method for SAR 3D reconstructed point cloud semantic segmentation.

Integrating Synthetic Aperture Radar (SAR) imaging with unmanned aerial vehicles (UAVs) plays a crucial role in urban area surveillance and situational awareness, benefiting from the low cost, small size, and high flexibility of SAR carried by drones. However, the dense arrangement of high-rise buildings in urban environments creates Urban Canyons and numerous visual blind spots due to occlusion, which weakens the perception capability of SAR. Additionally, SAR imaging results of moving vehicles on the roads between buildings result in severe defocusing due to their non-cooperative motion. In this paper, we establish a vehicle signal model for SAR imaging with UAVs that considers the vehicle body’s translation and the wheels’ rotation. The range history modulation and imaging characteristics of returns caused by translation and micro-motion are derived. Simulation results validate the correctness of the theoretical analysis, and the proposed theory helps explain SAR imaging results, providing support for high-precision focusing and three-dimensional imaging of non-line-of-sight (NLOS) SAR images.

Azimuth multichannel synthetic aperture radar (SAR) consistently encounters challenges associated with nonuniform sampling during its operation. The use of the conventional matrix inversion method leads to an increase in sampling irregularity, exacerbating noise interference during signal reconstruction and, consequently, contributing to a gradual decline in signal fidelity. This paper introduces a novel method to construct an L2regularized reconstruction matched filter, effectively mitigating noise impact prevalent in scenarios of extensive nonuniform sampling. The proposed method is highly computationally efficient, and the signal-to-noise ratio (SNR) index can be expressed in an unambiguous manner. Comparative simulation experiments illustrate the significant improvement in SNR performance achieved by the novel method in comparison to the traditional matrix inversion method.

Synthetic Aperture Radar Tomography (TomoSAR) has proven to be a powerful technique in urban mapping, disaster assessment, and counter-terrorism operations. However, existing inversion algorithms face challenges in balancing accuracy and efficiency. This paper introduces a novel gridless three-dimensional inversion method based on Weighted Least Squares (WLS) for TomoSAR, aiming to address issues related to accuracy, efficiency, and the lack of analytical expressions for the scatterer positions and backscattering coefficients. The validity of the proposed method is verified on the measured data.

In recent years, the geometric calibration for synthetic aperture radar (SAR) has been developing toward automation and lack of ground control points (GCPs). In this case, the calibration equations become ill-conditioned, making the results unstable. The slant range error of unmanned aerial vehicle (UAV)-borne SAR, whose transmitting signal is a frequency modulation continuous wave, is usually spatially variable, which makes this problem more complicated. In this letter, a geometric auto-calibration method for UAV-borne SAR is proposed. First, a robust auto-calibration model is proposed to calibrate the space-varying slant range error. Second, a weighting strategy based on the quality of tie points (TPs) is proposed to further improve the calibration accuracy of this method. Finally, an iterative method based on correcting characteristic values is utilized to find the least squares solution robustly. Simulation results confirm about 70% performance enhancement of the proposed method over the traditional approach. In real-data experiments, the auto-calibration yielded a positioning accuracy improvement to 11.5 cm.

A pulse repetition frequency (PRF) below the Doppler bandwidth is pivotal in reducing the costs of data transmission and increasing the swath width for synthetic aperture radar (SAR). However, azimuth undersampling leads to severe azimuth ambiguity. Existing methods for azimuth ambiguity suppression do not perform adequately under downsampling rate. To overcome the shortcomings of existing methods, this letter proposes an efficient deep unfolding network that combines group sparse and adaptive threshold techniques for SAR imaging, named group sparse and adaptive threshold techniques for SAR imaging (GSAT-Net). In GSAT-Net, thresholds and sparsity are incorporated for both the main area and ambiguity areas, allowing for threshold shrinkage that is not only layer-varied but also element-wise. Finally, the effectiveness of the proposed method is verified by visual comparisons and numerical analysis.

Synthetic aperture radar (SAR) tomography (TomoSAR) retrieves three-dimensional (3-D) information from multiple SAR images, effectively addresses the layover problem, and has become pivotal in urban mapping. Unmanned aerial vehicle (UAV) has gained popularity as a TomoSAR platform, offering distinct advantages such as the ability to achieve 3-D imaging in a single flight, cost-effectiveness, rapid deployment, and flexible trajectory planning. The evolution of compressed sensing (CS) has led to the widespread adoption of sparse reconstruction techniques in TomoSAR signal processing, with a focus on ℓ1norm regularization and other grid-based CS methods. However, the discretization of illuminated scene along elevation introduces modeling errors, resulting in reduced reconstruction accuracy, known as the “off-grid" effect. Recent advancements have introduced gridless CS algorithms to mitigate this issue. This paper presents an innovative gridless 3-D imaging framework tailored for UAV-borne TomoSAR. Capitalizing on the pulse repetition frequency (PRF) redundancy inherent in slow UAV platforms, a multiple measurement vectors (MMV) model is constructed to enhance noise immunity without compromising azimuth-range resolution. Given the sparsely placed array elements due to mounting platform constraints, an atomic norm soft thresholding algorithm is proposed for partially observed MMV, offering gridless reconstruction capability and super-resolution. An efficient alternative optimization algorithm is also employed to enhance computational efficiency. Validation of the proposed framework is achieved through computer simulations and flight experiments, affirming its efficacy in UAV-borne TomoSAR applications.

With the development of synthetic aperture radar (SAR) 3-D imaging technology, urban 3-D imaging has been realized by unmanned aerial vehicle (UAV)-borne array interferometric SAR (array-InSAR), which enables simpler access to SAR data. However, in low-altitude SAR 3-D imaging of urban scenes, multipath effect is significant. In past studies, multipath signals have been mostly considered as a hindrance in interpretation, but they provide important clues for imaging hidden targets in non-line-of-sight (NLOS) regions. Based on this idea, this article studies the multipath exploitation in low-altitude UAV-borne array-InSAR 3-D imaging and realizes accurate imaging of NLOS targets. An improved 3-D imaging model suitable for low-altitude cases is introduced to achieve preliminary 3-D imaging. Subsequently, key planes are extracted from the original results and the 3-D multipath reachable area is analyzed. After obtaining the multipath mechanism of the NLOS targets, they can eventually be accurately relocated and imaged. Both the simulation and real-data experiments show that the proposed method can effectively 3-D imaging and relocate NLOS targets in urban canyons, with errors typically below 0.5 m. The imageable range in the direction of the ground range can be expanded by 41.62%, and the imageable 3-D area can be expanded by 76.57%, which realizes the NLOS information acquisition.

Synthetic aperture radar (SAR) is a widely used active imaging system for remote sensing applications. However, traditional signal processing-based SAR imaging algorithms suffer from coherent speckle problems. Recently, statistical SAR imaging methods such as the FESAR model and plug-and-play (PnP) SAR imaging methods have been applied to suppress the speckle phenomenon. However, they are sometimes unstable and require an elaborate hyperparameter adjustment strategy during the iteration process. We propose extending PnP statistical imaging with Langevin sampling, called the Langevin-PnP algorithm. To construct the Langevin-PnP algorithm, we provide an in-depth analysis of the PnP framework and incorporate Langevin dynamics into its iteration trajectory. We also present a convergence guarantee for Langevin-PnP because the injected stochasticity affects the convergence condition under the law of probability. The experimental results showed that our proposed Langevin-PnP maintained the best performance over other statistical imaging methods, both in the simulated experiments and the RadarSat-SAR data experiments.

Breaking the constraint of pulse repetition frequency (PRF) is one of the important development trends of synthetic aperture radar (SAR). Within the conventional azimuth sampling patterns, severe ambiguity arises when confronted at a low PRF. Conversely, elevated PRF introduces considerable data redundancy, thereby culminating in wasting of resources. To address these issues, this paper proposes a novel joint optimization network for sparse SAR imaging and azimuth undersampling pattern grounded in the model-based reconstruction using deep learned priors (MoDL) architecture, combined with matched filter (MF) approximate measurement operators, named MF-based sampling pattern Joint optimization MoDL sparse SAR imaging Network (MF-JMoDL-Net). The MF-JMoDL-Net incorporates non-uniform sampling operators, enabling the sampling positions to be learnable, and achieves the groundbreaking joint optimization of the sampling pattern and ambiguity suppression. When the PRF is below the Nyquist sampling rate, the proposed network can acquire SAR images with minimal ambiguity and optimal imaging quality. Furthermore, the final learned undersampling pattern can be visualized and combined with the SAR echo signal semantics for mutual feedback. Extensive experiments on simulated and real scenes datasets are conducted to demonstrate the effectiveness and superiority of the proposed framework in imaging results.

Optical and synthetic aperture radar (SAR) image matching presents a formidable challenge due to their pronounced geometric and radiometric distinctions arising from multimodality. The distinct imaging mechanisms of optical and SAR sensors make it challenging to identify essentially homologous points in the physical sense, raising concerns about the accuracy and repeatability of correspondences in current feature matching methods. In this study, we introduce a detector-free feature matching algorithm specifically designed to match optical and SAR images through a two-step strategy. In the initial phase, our proposed method conducts pixelwise matching (PM) using downsampled feature descriptors, eliminating the necessity to identify repeatable keypoints. To mitigate complexity, we enforce a pseudo-epipolar constraint (PEC) to reduce computational costs by constraining the search range. Subsequently, refined matching is performed on the initial correspondences to rectify inaccuracies in the PM localization of the first step. Both matching steps are implemented on a graphics processing unit (GPU) to ensure high efficiency. The proposed algorithm attains an average matching accuracy of 2.39 pixels and operates with an efficiency of 1.09 s for 1108 image pairs, underscoring its superior comprehensive performance compared to various state-of-the-art algorithms, including handcrafted methods and deep learning networks.

Microwave photonic (MWP) SAR technology can realize large bandwidth even across multiple wavebands and therefore improves the imaging resolution significantly. However, when targets are illuminated by the electromagnetic wave with such an across-band bandwidth, their scattering characteristics will change with signal frequency, which is not taken into consideration for conventional SAR imaging. In this paper, in order to investigate the resolution and optimize the imaging of MWP SAR, we first design several typical structures and conduct electromagnetic simulations on them, from which, we obtain the variation of scattering characteristics within an across-band bandwidth. Then, we use different focusing approaches for one-dimensional imaging processing and adopt different evaluation indicators to evaluate the performance of these methods comprehensively. Finally, with detailed analysis, we give suggestions on the selection of compression methods for different structures at different SNRs.

Non-line-of-sight (NLOS) target imaging is a challenging issue in Synthetic Aperture Radar (SAR), because obstacles may block the direct line of sight between the radar and the target. In SAR tomographic (TomoSAR)3D imaging, the wrong presence of NLOS points can impact the accuracy and interpretability of the results. This paper presents an approach that not only detects but also relocates NLOS targets to their actual positions. The method relies on multipath signal analysis within a geometric model. The simulation and experimental results on a drone-borne system with 4 channels demonstrate that the proposed approach can effectively image NLOS targets, and has the potential for practical applications.

Tomographic SAR (TomoSAR) technology has gained significant attention in recent years due to its three-dimensional imaging capability. However, in practical applications, phase errors between different channels can degrade the quality of three-dimensional imaging. Current state-of-the-art methods for phase error compensation based on autofocus techniques suffer from high computational complexity, making them unsuitable for large-scale three-dimensional imaging. In this paper, we propose a multi-channel phase error estimation method based on error back-propagation training optimization. By utilizing the TomoSAR model that incorporates phase errors from multiple channels, we construct a matrix containing the parameters to be estimated for inter-channel phase errors. Through stochastic gradient descent algorithm, we iteratively optimize the parameters of the phase error matrix, ultimately obtaining an estimation of the inter-channel phase errors. Experimental results validate the accuracy of the proposed method.

SAR tomography (TomoSAR) has the three-dimensional (3-D) resolving ability. Most existing 3D imaging methods for TomoSAR consider the layovers for different channels are the same. However, in the low-altitude airborne cases, the variation of layovers between channels becomes unneglectable. In this letter, we studied the channel migration phenomenon of sparse TomoSAR under low-altitude scenarios. We derived the model of the differential range from a particular target to different antennas, which is nearly a linear function along the array dimension. Therefore, we applied Keystone Transform on the array axis to correct this channel migration, and then the traditional compressed sensing-based 3D imaging methods for TomoSAR can be applied to get the final 3D reconstruction results. The approach was tested with simulation data and also the real data of the MV3DSAR system, which is a mini drone-borne TomoSAR system. The results demonstrate that the proposed method can provide a more accurate solution to the low-altitude sparse TomoSAR reconstruction problem.

Edge detection is a technique used to identify inherent structures within an image, and it is an essential requirement for synthetic aperture radar (SAR) applications. In particular, ratio-based edge detectors have been widely used in SAR image registration because of their ability to extract invariant features and reduce the effects of speckle noise. However, current edge detectors often struggle to accurately detect multi-scale objects and low-contrast structures. To address this issue, we present a robust multi-scale edge detector that uses a modified convolution kernel to improve the extensibility of edge features and aggregates multi-scale feature responses. We also propose a local scale estimation module to enhance edge responses in low-contrast areas and reduce noise effects. The experimental results demonstrate that our proposed method effectively preserves the integrity, continuity, and robustness of multi-scale and low-contrast structures. By incorporating our proposed edge detector into feature and template matching frameworks, we are able to significantly improve matching accuracy and outperform state-of-the-art SAR image registration methods.

Ship target recognition is a vital task in synthetic aperture radar (SAR) imaging applications. Although convolutional neural networks have been successfully employed for SAR image target recognition, surpassing traditional algorithms, most existing research concentrates on the amplitude domain and neglects the essential phase information. Furthermore, several complex-valued neural networks utilize average pooling to achieve full complex values, resulting in suboptimal performance. To address these concerns, this paper introduces a Complex-valued Convolutional Neural Network (CVGG-Net) specifically designed for SAR image ship recognition. CVGG-Net effectively leverages both the amplitude and phase information in complex-valued SAR data. Additionally, this study examines the impact of various widely-used complex activation functions on network performance and presents a novel complex max-pooling method, called Complex Area Max-Pooling. Experimental results from two measured SAR datasets demonstrate that the proposed algorithm outperforms conventional real-valued convolutional neural networks. The proposed framework is validated on several SAR datasets.

Convolutional neural networks (CNNs) are widely used in image classification, but such methods often require massive labeled data as learning resources. On the one hand, synthetic aperture radar (SAR) image interpretation is difficult, resulting in the lack of large-scale data sets with high-quality labels. On the other hand, CNNs are not explainable enough to provide reliable and trusted application services for SAR target recognition. To solve the above problems, physics-based electromagnetic simulated images are used to alleviate the shortage of real data with annotations, and explainability analysis methods are introduced to explain the basis of network decision-making. To address the domain gap between simulated and measured data, we propose a novel network integrating dual-branch image reconstruction and subdomain alignment (DBRSA). The network completes the reconstruction of simulated and measured images through the domain-shared encoder and domain-specific decoders, thereby helping the encoder to learn feature extraction methods independent of labels. In addition, the network aligns the feature vectors of similar targets obtained from different domains according to the real or pseudo labels of the samples, so as to further improve the classification accuracy. The experimental results and model decision analysis results demonstrate that the proposed network can improve the performance reliably by reducing the attention to the background noise and increasing the attention to the shadows and contours of the target, effectively reducing the dependence on the number of sample labels in practical application scenarios.

This paper presents a novel efficient method for gridless line spectrum estimation problem with single snapshot and sparse signals, namely the gradient descent least squares (GDLS) method. Conventional single snapshot (a.k.a. single measure vector or SMV) line spectrum estimation methods either rely on smoothing techniques that sacrificing the range and/or azimuth resolution, or adopt the sparsity constraint and utilize compressed sensing (CS) method by defining prior grids and resulting in the off-grid problem. Recently emerged atomic norm minimization (ANM) methods achieved gridless SMV line spectrum estimation, but its computational complexity is extremely high; thus it is practically infeasible in real applications with large problem scales. Our proposed GDLS method reformulates the line spectrum estimations problem into a least squares (LS) estimation problem and solves the corresponding objective function via gradient descent algorithm in an iterative fashion with efficiency. The convergence guarantee, computational complexity, as well as performance analysis for evenly distributed antenna array case are discussed in this paper. Numerical simulations show that the proposed GDLS algorithm outperforms the state-of-the-art methods e.g., CS and ANM, in terms of estimation performances. It can completely avoid the off-grid problem, and its computational complexity is significantly lower than ANM. Our method has been tested in tomographic SAR (TomoSAR) imaging applications via simulated and real experiment data. Results show great potential of the proposed method in terms of better cloud point performance and eliminating the gridding effect.

Tomographic SAR technique has attracted remarkable interest for its ability of three-dimensional resolving along the elevation direction via a stack of SAR images collected from different cross-track angles. The emerged compressed sensing (CS)-based algorithms have been introduced into TomoSAR considering its super-resolution ability with limited samples. However, the conventional CS-based methods suffer from several drawbacks, including weak noise resistance, high computational complexity, and complex parameter fine-tuning. Aiming at efficient TomoSAR imaging, this paper proposes a novel and efficient sparse unfolding network based on the analytic learned iterative shrinkage thresholding algorithm (ALISTA) architecture with adaptive threshold, named Adaptive Threshold ALISTA-based Sparse Imaging Network (ATASI-Net). The weight matrix in each layer of ATASI-Net is pre-calculated as the solution of an off-line optimization problem, leaving only two scalar parameters to be learned from data, which significantly simplifies the training stage. Furthermore, the introduction of an adaptive threshold for each azimuth-range pixel permits the threshold shrinkage to be not only layer-varied but also element-wise. Additionally, the final learned thresholds can be visualized and combined with the SAR image semantics for mutual feedback. Finally, extensive experiments on simulated and real data are carried out to demonstrate the effectiveness and efficiency of the proposed method.

Multi-sensor remote sensing applications require the registration of optical and Synthetic Aperture Radar (SAR) images, which presents challenges due to significant radiometric and geometric differences resulting from distinct imaging mechanisms. Although various algorithms have been proposed, including hand-crafted features and deep learning networks, most of them focus on matching radiometric-invariant features while ignoring geometric differences. Furthermore, these algorithms often achieve promising results on datasets that use manually labeled ground truths that may be less reliable for high-resolution SAR images affected by speckle noise. To address these issues, we propose a robust global-to-local registration algorithm consisting of four modules: geocoding, global matching, local matching, and refinement. We generate a geometry-invariant mask in the geocoding module to help the local matching module focus on valid areas, introduce a fast global matching method to solve large offsets, and use matching confidence to guide subsequent local matching based on the accuracy of global matching. We propose a feature based on multi-directional anisotropic Gaussian derivatives (MAGD) and embed it into the confidence-aware local matching with the geometry-invariant mask to reduce the effect of geometric differences. Finally, we refine correspondence positions and remove outliers. We also build a high-accuracy evaluation dataset with hundreds of image pairs, where the ground truth is obtained by meta poles, which have clear and reliable structures in both optical and SAR images. Experimental results on this dataset demonstrate the superiority of our proposed algorithm compared to several state-of-the-art methods.

In order to explore a better classification strategy for multi-band polarimetric SAR, we apply both the deep learning classifier and the clustering method based on scattering mechanism to the airborne polarimetric SAR data of P, L, S,$\mathrm{C}$, X band. Firstly, the characteristics of deeplab$\mathrm{v}3+$segmentation results in different frequency bands are analyzed. Then, the performance of the clustering results based on the h-alpha-wishart method in different frequency bands is analyzed. Through the comparison of classification results, we summarize the characteristics of different frequency bands, which provides support for subsequent classification strategy optimization.

An innovative approach, mainly featured by progressive iterations of partitioning and focusing, to processing very-high-resolution spaceborne synthetic aperture radar (SAR) data is presented in this article. Due to long integration time endured during data acquisition as well as large scene extension being common to the advanced SAR systems, spatial dependence of range histories may be severe, especially is the terrain undulation severe, and must be properly dealt with. In this article, after deriving the exact two dimensional spectrum for an curved satellite orbit, we focus the entire echo data through several iterations of partitioning and focusing. Besides, a new technique aiming at enhance the focusing quality of the final data blocks is also presented. Finally, the well-focused final image blocks are recombined into the entire image. Owing to the feature of partitioning, the spatial dependence of many factors can be easily accommodated. The methodology is validated by processing results on simulated and real SAR data.

In recent years, the stereo positioning technology based on multiview spaceborne synthetic aperture radar (SAR) images has been widely applied in digital surface model extraction. In this letter, problems of the existing methods based on the range–Doppler (RD) model in a multiview stereo solution are pointed out. A robust stereo positioning solution for multiview spaceborne SAR images based on the RD model is proposed. In the proposed method, the traditional RD model is normalized to reduce the model errors caused by the different scales of the range equation and Doppler equation. A weighting strategy is also proposed to improve the positioning accuracy. This strategy is useful for multiview stereo positioning when using images of different satellites with orbital data of different accuracies. The experiments based on GaoFen-3 and TerraSAR-X satellite data sets validate the effectiveness of the proposed method.

Unlike geocoded images, slant-range (SR) synthetic aperture radar (SAR) images vary from imaging resolution to angles, which are difficult to be registered directly using the traditional SAR image registration methods. A possible way is to match their corresponding geocoded images and to project the correspondences to SR images. However, this way is time consuming and suffers from both registration and projection errors. In this letter, an automatic and efficient method is proposed to directly match multiview SR SAR images. We first estimate the scale and rotation differences between two SR images from the metadata delivered by vendors alongside the image file. Specifically, the scale differences of the range and azimuth directions are estimated by transforming the range and azimuth pixel intervals into a uniform geographical resolution, and the rotation differences are estimated by comparing the azimuth angles of an image-pair. A global-to-local framework is then implemented to accelerate the registration process. In the global stage, we fix the scale and rotation parameters in SAR-scale-invariant-feature-transform (SAR-SIFT) method to avoid mismatches. In the local stage, the phase correlation of cropped patches is parallelized to generate accurate matches. Experimental results on 13 multiview SAR images of the Omaha city show that the proposed method can provide accurate and efficient registration results for each pair of the 13 images, and outperforms the state-of-the-art methods both in accuracy and in efficiency.

Building segmentation of synthetic aperture radar (SAR) images is a challenging task that has not been solved well. High-resolution interferometric SAR (InSAR) images can provide delicate textures and interferometric phase images useful for building segmentation. However, current semantic segmentation networks in computer vision cannot be directly applied in InSAR building segmentation tasks to get good results because of the InSAR images’ particularity. In this article, we present a novel complex-valued convolutional and multifeature fusion network (CVCMFF Net) specifically for building semantic segmentation of InSAR images. This CVCMFF Net not only learns from the complex-valued SAR images but also considers multiscale and multichannel feature fusion. It can effectively segment the layover, shadow, and background on both the simulated InSAR building images and the real airborne InSAR images. The segmentation performance of CVCMFF Net is significantly improved compared with those of other state-of-the-art networks. By feature visualization, the feature extraction rule and feature fusion mechanism of the network are explored. We hope that the proposed network can be beneficial to InSAR phase filtering, phase unwrapping, and information extraction in urban areas.

In airborne array interferometric SAR (Array-InSAR) tomography, the measurements acquired by conventional uniform sampling array are always restricted by the number of physical baseline elements and the size of baseline aperture. It is desirable to capture new acquisitions and enlarge the aperture with virtual signal processing instead of actually adding array baselines. For this motivation, we utilize the disparity of a pair of coprime sampling sub-arrays to enlarge the baseline aperture and construct new observations virtually. The generation of virtual measurements is equal to estimating cross-correlation matrices in real SAR data. Due to the spatial target variation, we adopted an adaptive filtering method to estimate the cross-correlation matrix. We call the above-mentioned processing of generating virtual measurements as acoprime sensing technique. The newly generated virtual measurements have more degrees of freedom, a larger baseline aperture, and a higher signal-to-noise ratio (SNR) than the physical measurements. These advantages offer the possibility to obtain competitive three-dimensional (3-D) imaging results without increasing the hardware cost of the Array-InSAR. We demonstrate the effectiveness of the proposed method by the coprime acquisitions selected from AIRCAS Array-InSAR data.

Despite impressive progress in the past decade, accurate and efficient multiview synthetic aperture radar (SAR) image registration remains a challenging task due to complex imaging mechanisms and various imaging conditions. Especially, for rugged areas, SAR images obtained from the opposite-side view reflect different characteristics, making popular SAR image registration methods no longer applicable. To this end, we propose a geometry-aware image registration method by extracting inherent orientation features and concentrating on geometry-invariant areas. First, slant range images are terrain-corrected using a digital elevation model (DEM) to reduce large relative positioning errors caused by elevation. Second, the Gabor-ratio detector is introduced to obtain multiscale orientation features, which are more robust under various imaging conditions. Then, a geometry-aware mask is produced by intersecting the 3-D space ray with DEM, and thus, SAR images can be divided into three categories, layover, shadow, and geometry-invariant areas. The geometry-aware matching method, which focuses on geometry-invariant areas and masks out misleading caused by geometric and radiometric distortions, is proposed to realize accurate matching. The rational polynomial coefficients (RPCs) are refined to achieve relative correction. Extensive results on dozens of SAR images demonstrate the effectiveness and universality of the proposed algorithm by quantitative evaluation using man-made and natural corner reflectors. An analysis of the factors affecting registration accuracy is also discussed.

Unsupervised methods play an essential role in Polarimetric SAR (PolSAR) image classification, where labeled data are difficult to obtain. However, there is still a large gap between existing unsupervised learning methods and supervised learning methods. Without the semantic constraints of labeled data, pixels within the same category are often misclassified into different categories, leaving the output to be messy. To address the previous issue, this paper proposes a fully unsupervised pipeline for training CNNs. The pipeline combines low-level superpixels and high-level CNN semantic features for high-quality pseudo label generation. It effectively eliminates the misclassified pixels by voting within the superpixel blob, while preserving the sharpness of edges. With the training process of the model, the quality of the generated labels is getting improved. Experiments on airborne (ESAR/AIRSAR) and spaceborne (RadarSat2) PolSAR images prove the effectiveness of the proposed method (measured with OA, AA, and Kappa metrics). Our method outperforms the previous unsupervised methods (H/alpha-Wishart,SM-Wishart, FDD-H, DEC, and VQC-CAE) with a large margin, and even has comparable performance to the supervised CNN model (FCN).

Deep learning algorithms are widely used in SAR target detection. At present, most detection methods based on neural networks treat SAR images as optical images for processing, and do not fully exploit the characteristics of SAR images. In this paper, a method for feature enhancement using sub-aperture decomposition is proposed. Considering the advantages of the rotate box in eliminating interference, we conducted experiments on the rotating frame detection network Rotate RetinaNet using the Complex SAR images Rotation Ship Detection Dataset (CSRSDD). Compared with the base method, AP improves by 2.4% without bells and whistles. The results confirm the effectiveness of the proposed method.

High-resolution (HR) SAR images contain rich details thus offer better capabilities on describing the target spatial properties. Besides, Polarimetric SAR (PolSAR) data can convey physical characteristics embedded in the polarimetric information. However, HR and polarimetric are two conflicting properties since acquiring PolSAR data will limit the pulse repetition frequency (PRF) for each polarization hence limit the resolution. Tackling this issue, recovering HR full-polarized (FP) data from HR single-polarized (SP) data using DNN is a meaningful approach. This paper studies the ability of DNN to recover FP information from SP data of different resolutions and different polarization channels in various areas. We use UAVSAR L-band data for experiments. Results demonstrate the feasibility of using DNN trained by lower-resolution FP data to recover HR FP data from HR SP data. It is showed that the performance is more notable over natural surfaces than in built-up areas, and HH polarization has a better reconstruction result compared with other channels. Besides, the recovery capability in different regions is also compared, which provides some inspirations for FP and HR SAR data applications.

Three-dimensional imaging with tomographic SAR is a hot research topic in the field of SAR. The existing methods usually solve the question based on elevation discretization. When the target point has an off-grid deviation, the accuracy will decrease and the number of spurious points will increase. In this paper, a position-first step-by-step 3D inversion method is proposed. By constructing the observation equation with only an unknown elevation position, the height of the scattering center is directly solved in the continuous domain based on the least square criterion, and then the scattering coefficient is calculated based on the position. Simulation results verify the effectiveness of the method compared with the classical OMP algorithm. The proposed method has higher estimation accuracy under the same noise level and observation number and is less affected by the scattering center spacing.

The super-resolution algorithm can break through the theoretical resolution limit of radar and improve image resolution by several times under the condition of hardware parameter limitation. The improved resolution can see more details of the image, which is helpful for target detection and recognition. In this paper, a 2D spatial smoothing music superresolution FMCW SAR imaging algorithm is proposed. In this algorithm, the spatial smoothing MUSIC algorithm is used to establish the guidance vector relationship between the frequency domain and the spatial position of the targets, and GDE is used to estimate the number of sources. The algorithm can achieve the super-resolution reconstruction of the imaging target space.

Understanding the physical properties and scattering mechanisms contributes to synthetic aperture radar (SAR) image interpretation. For single-polarized SAR data, however, it is difficult to extract the physical scattering mechanisms due to lack of polarimetric information. Time-frequency analysis (TFA) on complex-valued SAR image provides extra information in frequency perspective beyond the “image” domain. Based on TFA theory, we propose to generate the subband scattering pattern for every object in complex-valued SAR image as the physical property representation, which reveals backscattering variations along slant-range and azimuth directions. In order to discover the inherent patterns and generate a scattering classification map from single-polarized SAR image, an unsupervised hierarchical deep embedding clustering (HDEC) algorithm based on TFA (HDEC-TFA) is proposed to learn the embedded features and cluster centers simultaneously and hierarchically. The polarimetric analysis result for quad-pol SAR images is applied as reference data of physical scattering mechanisms. In order to compare the scattering classification map obtained from single-polarized SAR data with the physical scattering mechanism result from full-polarized SAR, and to explore the relationship and similarity between them in a quantitative way, an information theory based evaluation method is proposed. We take Gaofen-3 quad-polarized SAR data for experiments, and the results and discussions demonstrate that the proposed method is able to learn valuable scattering properties from single-polarization complex-valued SAR data, and to extract some specific targets as well as polarimetric analysis. At last, we give a promising prospect to future applications.

This article proposes a novel synthetic aperture radar visual-inertial odometry (SAR-VIO) consisting of an SAR and an inertial measurement unit (IMU), which aims to enable the observation platform to complete successfully a continuous observation mission in the context of low-cost demand and lack of enough navigation information. First, we establish the observation models of the SAR in a continuous observation process based on the SAR frequency-domain imaging algorithm and the SAR time-domain imaging algorithm, respectively. With the preintegrated IMU data, we then propose a method for estimating the geographic locations of the matched targets in the SAR images and verify the condition and correctness of the method. The optimization of the track and the locations of the targets is achieved by bundle adjustment according to the minimum reprojection error criterion, and a sparse point-cloud map can be obtained. Finally, these methods and models are organized into a complete SAR-VIO framework, and the feasibility of the framework is verified through experiments.

Synthetic aperture radar (SAR) tomography (TomoSAR) systems have been widely used because of its capability of 3D reconstruction of urban area. However, due to the critical requirements of phase and amplitude accuracy, the channel imbalance of the airborne multi-baseline TomoSAR system must be estimated and compensated. In this paper, we proposed a channel imbalance calibration method with visual semantics. This method only uses the information from 2D SAR images and works well in the absence of ground control points (GCPs). Firstly, strong scatterer in 2D image is selected and the channel imbalances are estimated based on the SAR imaging geometry. Secondly, altitude of the scatterer is estimated with TomoSAR technique and the imaging geometry can be updated accordingly. After some iterations, the channel imbalance will converge and then be compensated. The estimated error is compared with the results based on GCPs and 3D point cloud are presented, which validates the feasibility of proposed method.

As the first C-band full-polarization spaceborne synthetic aperture radar (SAR) satellite in China, Gaofen-3 (GF-3) is of great significance in scientific research and economic development. Thanks to the advantage of all-weather and all-day observation, spaceborne SAR imagery is widely used in digital elevation model (DEM) generation, target tracking, and so on. And, all these applications are based on the high 3-D geolocation accuracy of SAR imagery. Therefore, a new slant range error update model and an error-source-based weight strategy which can effectively improve the 3-D geolocation accuracy of multiobserved data set from GF-3 satellite are proposed in this letter. The slant range error update model is proposed to improve the 3-D geolocation accuracy using the virtual control information in height direction. At the same time, an error-source-based weight strategy is introduced as an appendant to the multiobservation block adjustment model based on the rational function model (RFM). The experimental results demonstrate that our proposed method can achieve a much higher 3-D geolocation accuracy than the traditional methods.

Due to multipath (MP) scattering, many special phenomena and details about the targets in high-resolution synthetic aperture radar (SAR) images are hard to understand. In this letter, a more elaborate improved model combined with the synthetic aperture progress is presented, to analyze the MP scattering. We explore the defocusing and displacement phenomena of MP scattering through signal modeling and time-frequency relationship analysis. The theoretical equations for the displacement of double scattering are derived, and the equation of phase error which causes the defocusing is also given. The SAR raw data simulation of MP scattering and range-Doppler (RD) imaging results validates the correctness of the analysis. From the equations, the effects of MP scattering of different height and different orientation angle in high-resolution SAR images are given. Finally, based on the simulation results, the new interpretation about the relationship between the moustache effect in high-resolution TerraSAR-X SAR images and the height of the structure is obtained.

Synthetic aperture radar (SAR) image simulation can provide SAR target images under different scenes and imaging conditions at a low cost. These simulation images can be applied to SAR target recognition, image interpretation, 3-D reconstruction, and many other fields. With the accumulation of high-resolution SAR images of targets under different imaging conditions, the simulation process should be benefited from these real images. Accurate simulation parameters are one of the keys to obtain high-quality simulation images. However, it takes a lot of time, energy, and resources to get simulation parameters from actual target measurement or adjusting manually. It is difficult to derive the analytical form of the relation between a SAR image and its simulation parameter, so nowadays the abundant real SAR images can hardly help the SAR simulation. In this article, a framework is proposed to obtain the relationship between SAR images and simulation parameters by training the deep neural network (DNN), so as to extract the simulation parameters from the real SAR image. Two DNNs, convolutional neural network (CNN), and generative adversarial network (GAN) are used to implement this framework. By modifying the network structures and setting reasonable training data, our DNNs can learn the relationship between image and simulation parameters more effectively. Experimental results show that the DNNs can extract the simulation parameters from the real SAR image, which can further improve the similarity of the simulation image while automating the setting of simulation parameters. Compared with CNN, the simulation parameters extracted by GAN can achieve better results at multiple azimuth angles.

The GF-3 satellite is the first full-polarization SAR satellite in China, which operates in the C band with a resolution of 1m. This paper presents a dual-aspect geometric terrain correction method to overcome the inherent shortages of SAR image such as foreshortening, shadow and layover. The results using GF-3 satellite data show that the method can effectively eliminate layover and shadow distortions in SAR images. This method solves the geometric correction problem that cannot be solved with a single SAR image．

Obtaining the three-dimensional data of the target is very useful for the interpretation and application of the SAR target. This paper proposes a deep learning framework to recover the three-dimensional structure of the target from a single SAR image, which is expressed in the form of 3D point cloud. Due to the small data set of SAR images, the network is combined by two parts. First, the two-dimensional image in the optical perspective is predicted from the SAR target image, and then the 3D points of the target is reconstructed based on the pre-trained 3D reconstruction network model from the optical images. The experiment is based on the MSTAR datasets. The results confirm the effectiveness of the three-dimensional reconstruction method.

With the needs of quality assessment for massive GF-3 polarimetric data, a method based on common distribution targets has been proposed by Sha Jiang. However, it needs manually selection of those woodlands, and cannot be performed automatically. In this paper, an automated GF-3 full-polarization SAR data quality assessment method is conducted using a classic Convolution Neural Network (VGG-16). The network is pre-trained by Radarsat-2 PolSAR data and then trained by selected typical GF-3 scenes. It is supposed to learn the features of the targets, which satisfies the azimuthal symmetry and backscatter reciprocity and fulfills the quality assessment work. Several typical GF-3 strips data are used to test the method. Experiments show that the network can predict the plots of targets from a new scene under the interference of polarimetric distortion and noise. And, the quality assessment results by the network are consistent with the manual assessment results, which shows the effectiveness of the method.

High-resolution SAR has large transmitting bandwidth and wide synthetic aperture. How to understand and take advantage of the variation characteristics of SAR scattering characteristics with angle and frequency is a topic that worth studying. This article establishes a coherence matrix of sub-band and sub-aperture SAR images, and analyzes its ability to classify scattering mechanism. Experiments are conducted using the TerraSAR-X high-resolution data of different scenarios, and some meaningful results are got, which may provide some support to the analysis and application of high-resolution SAR data.

The dominant scattering mechanism is of great significance for the application of ground objects classification and target detection. It can also verify the quality of the polarimetric data by check the dominant scattering mechanism of known ground objects. In order to improve the application performance, this paper studies the dominant scattering mechanism of GF-3 typical ground objects based on a large number of data slices. The GF-3 fully polarimetric data slices are classified based on the MODIS global classification map, and the GF-3 slice library of typical ground objects is constructed. Based on large amounts of GF-3 samples, we carry out the statistical analysis of dominant scattering mechanism separation results for typical GF-3 ground objects (building, woodland, cultivated land, grassland and waters) of by means of h/alpha/A decomposition. The quantitative results reveal the polarimetric scattering feature of different ground objects, and provide reference for fully polarimetric SAR application.

Accumulation of target features from multi-angle observation data is of great significance for SAR Automatic Target Recognition. How to make correspondence matching with multi-angle data effectively for man-made targets? Due to the intensities and positions of the scattering centers varies with the aspect angle strongly, it is difficult to register SAR images using traditional matching methods. In this work, an approach for feature matching of man-made targets is proposed from complex SAR data. And it is based on the mechanisms of the dominate scattering to build up the relationship of the different multi-aspect images. Scattering map is employed to make scattering center extractions. Then the scattering centers are matching by using the physical model. Terra-SAR data of an aircraft test the validity of the approach.

Classification of intertidal area in synthetic aperture radar (SAR) images is an important yet challenging issue when considering the complicatedly and dramatically changing features of tidal fluctuation. The difficulty of intertidal area classification is compounded because a high proportion of this area is frequently flooded by water, making statistical modeling methods with spatial contextual information often ineffective. Because polarimetric entropy and anisotropy play significant roles in characterizing intertidal areas, in this paper we propose a novel unsupervised contextual classification algorithm. The key point of the method is to combine the generalized extreme value (GEV) statistical model of the polarization features and the Markov random field (MRF) for contextual smoothing. A goodness-of-fit test is added to determine the significance of the components of the statistical model. The final classification results are obtained by effectively combining the results of polarimetric entropy and anisotropy. Experimental results of the polarimetric data obtained by the Chinese Gaofen-3 SAR satellite demonstrate the feasibility and superiority of the proposed classification algorithm.

A fully automatic, non-Gaussian, and contextual clustering algorithm for segmentation of polarimetric synthetic aperture radar (SAR) images has been previously presented by Doulgeris. It achieved good results for both simulated and actual data sets. However, the long computation time was its main drawback. This letter discusses modifications to improve computational efficiency. The primary speed issues were rooted in the complicated probability density function (PDF) of the adopted model, for which evaluating the posterior probability of samples and estimating the parameters were both very time-consuming. We investigate the model parameters, reparametrize the model, and introduce lookup tables to speed up the processing chain. The new strategy speeds up both PDF evaluation and parameter estimation while maintaining the exactly similar visual results and now makes advanced non-Gaussian SAR image analysis a practical alternative.

The stability characteristics of scattering are of great significance for achieving radiometric calibration without calibration fields. This letter used dual-polarization synthetic aperture radar (SAR) data to explore the temporal and spatial stability of microwave scattering. Based on Sentinel-1 SAR data, we mainly studied the stability of the median value of SAR image slices in urban areas, under the HH- and HV-polarization modes. The urban slices with stable scattering were extracted by an improved deep neural network designed herein. Based on 33 images of the Houston region, the results show that this suitable network, which is trained by dual-polarization data, can effectively distinguish the stable slices from the unstable ones. In addition, some meaningful characteristics about SAR scattering stability of dual-polarization data are also obtained, which can provide a good reference for SAR nonfield radiometric calibration.

Target classification is an important part in automatic target recognition (ATR) systems. Deep learning methods get state of the art performance in SAR target classification. Simulation is a useful data augmentation method when the numbers of real samples for training is not sufficient. This article discusses how to release the full potential of simulated samples which is used to improve performance of SAR target classifier. The proposed method is based on cycle adversarial network (CycleGAN), which can transfer simulated samples to be more similar with real samples in image domain. Experiments show that adding simulated samples straightforward into training dataset is not helpful to improve the performance. However, adding the transferred simulated samples for training results in about 10% increase in accuracy in the designed SAR airplane classification experiment, compared with training without data augmentation.

The theoretical modeling and analysis of SAR location error play an important role in SAR system design and error source budget. Existing SAR geolocation error models are mainly implicit, which are not easy to do analysis, especially in the curved-path case. In this paper, a theoretical explicit model of the relationship between image geolocation error and the path measurement error is established for curved-path SAR, based on BP imaging algorithm. Simulations are given which verify the correctness of the model. The explicit model and the analysis results provide an effective reference for understanding and budgeting the system-level geometric location error for curved-path SAR, such as GeoSAR.

Scalloping is one of the critical problems in ScanSAR images. It not only affects image visualization, but also influences the quantitative applications such as surface wind and wave retrievals in the ocean area. The existing method of descalloping needs artificial parameter setting and lacks generality in the image domain. A novel deep neural network based on residual learning for descalloping of ScanSAR images is proposed in this paper. The proposed method can eliminate scalloping patterns and has strong adaptive ability, which can handle inhomogeneous scalloping patterns and different scenarios. Experiments on GF-3 ScanSAR images verify the good performance of this method. The code for our models is available online.

SAR images of aircrafts usually consist of several discrete scattering centers. For targets like aircrafts having relatively smooth surfaces, understanding and modeling about these scattering centers is of great use for the detection and the recognition in SAR images. In this work, an approach for modeling of the features about the dominant scattering centers of the aircrafts is proposed. The approach generates a scattering map based on the analytic scattering model of some typical unit structures. According to scattering maps, the equivalent parameters are adjusted for detail feature modeling. Through this approach, the main mechanisms of aircrafts can be interpreted and the main features are modeling too. This approach provides a way to characterize the features of aircrafts in SAR images without the large computational amount. Furthermore, it sets up a frame based on typical structures to model the local scattering features of the aircrafts in SAR images. The experiments based on the Terra-SAR data and airborne data test the validity of the approach. The work would be useful for SAR Automatic Target Recognition.

A new image formation algorithm is proposed for processing very high resolution spaceborne sliding-spotlight synthetic aperture radar (SAR) data. Because of along-track antenna steering, the Doppler bandwidth of the received SAR data is expanded significantly beyond one pulse repetition frequency interval. Furthermore, the range histories become spatially dependent in both dimensions and cannot be expressed exactly by a hyperbolic model. In our approach, we first reduce the Doppler bandwidth by a novel azimuth dechirp processing method in the range frequency domain. The data are then processed by the standard ω-κ algorithm with a fixed effective velocity. Thereafter, the chirp modulation concept is imported to rebuild new data with much shorter apertures. Finally, a standard back-projection algorithm is employed to accumulate the signal pixel by pixel along the newly built aperture. Thus, the balance between processing efficiency and precision can be controlled by adjusting the length of the new apertures. In addition, a more accurate 2-D spectrum derivation is employed to enhance the processing precision, and a novel range-splitting method is presented to accommodate the range dependence of effective velocities. Furthermore, when implementing the back projection, the image grid-the region and granularity level of which are user defined-is placed on the earth's surface instead of on the slant-range plane, and the routine geometry projection processing thus becomes dispensable.

GaoFen-3 is China's first meter-level multi-polarization Synthetic Aperture Radar (SAR) satellite with scientific and commercial applications, which was developed by the China Academy of Space Technology (CAST) and had been launched in August, 2016. The SAR instrument and ground data processing system were developed by the Institute of Electronics, Chinese Academy of Sciences (IECAS). It employs a multi-polarization C-band SAR based on active phased array technology, which allows flexible beam operations in azimuth scanning, range scanning, right looking and left looking. Hence, GaoFen-3 has 12 imaging modes, covering the traditional Stripmap mode, ScanSAR mode, and the emerging Wave mode and Sliding Spotlight mode, and is a SAR satellite with the most abundant imaging modes in the world. In order to evaluate the imaging performance of these modes, the multiple mode SAR raw data simulation is highly demanded. In the paper, the simulation framework, the simulation algorithms and the evaluation strategies will be briefly introduced to expose how the raw data simulation guarantees the development of GaoFen-3 and its processing system.

Target recognition in synthetic aperture radar (SAR) images has become a hotspot in recent years. The backscattering characteristic of target is a significant issue taken into consideration in SAR applications. Almost all of the previous work focus on the scatter point extraction to depict the backscattering characteristic of the target; however, a point-target corresponds to a region rather than a single point due to the convolution during the imaging. Based on this fact, we first analyze the extent to how a point-target spreads, then propose a novel scatter cluster extraction (SCE) method, and utilize the scatter cluster as the feature to solve the airplane recognition problem in SAR images. In practice, there often exist interfering objects near the target to be classified. To overcome this issue, we design a reweighted sparse representation (RSR)-based automatic purifying method by assigning a weight to each element of the feature iteratively according to the representation error. Since the element with large representation error always corresponds to the interfering objects, we give it a small weight, consequently suppressing the influence of the interference. Experimental results demonstrate that the proposed SCE method outperforms the traditional scatter point extraction-based method as well as some state-of-the-art methods. The comparison result also validates the effectiveness of the proposed RSR method.

Ship target recognition has always been a hot issue in the field of ocean surveillance. Due to the serious shortage of samples in ship target recognition for synthetic aperture radar (SAR) images, the template-based method is still one of the most effective ways to solve the problem. In this letter, we put forward a novel ship recognition method based on the projection shape template (PST), aiming at increasing both the accuracy and the robustness of the recognition. The PST of each category is calculated by projecting the 3-D model obtained from the two-view images of the target to the 2-D slant-plane image according to the SAR imaging model. Then, we propose a contour extraction method to detect the profile of ships, which served as the feature. Finally, the identity of the query ship is obtained through contour matching. Experimental results indicate that the proposed method is effective even when the number of samples is extremely small, consequently providing a promising way for the automatic interpretation of ship targets in the SAR images.

This paper addresses the impact of mixtures between classes on equivalent number of looks (ENL) estimation. We propose an unsupervised ENL estimator for polarimetric synthetic aperture radar (PolSAR) data, which is based on small sample estimates but incorporates a mixture-eliminating (ME) procedure to automatically assess the uniformity of the estimation windows. A statistical feature derived from a combination of linear and logarithmic moments is investigated and adopted in the procedure, as it has different mean values for samples from uniform and nonuniform windows. We introduce an approach to extract the approximated sampling distribution of this test statistic for uniform windows. Then the detection is conducted by a hypothesis test with adaptive thresholds determined by a nonuniformity ratio. Finally the experiments are performed on both simulated and real SAR data. The capability of the unsupervised ME procedure is verified with simulated data. In the real data experiments, the ENL estimates of Flevoland and San Francisco PolSAR images are analyzed, which show the robustness of the proposed ENL estimation for SAR scenes with different complexities.

A novel unsupervised, non-Guassian and contextual clustering algorithm for segmentation of polarimetric SAR images has been presented in [1]. This represents one of the most advanced PolSAR unsupervised statistical segmentation algorithm and uses the doubly flexible, two parameter, U-distribution model for the PolSAR statistics. However complexity of the probability density function leads to high time consumption. This paper investigate the key dependent variable in the U-distribution model and find a new parameter domain where the PDFs are smooth. Then a one-dimensional look-up table is set in this domain with nodes number determined by corresponding Fourier spectrum and is adopted to avoid re-evaluating the numerical integral in PDF to calculate class posteriori probabilities for every sample. The proposed strategy is incorporated in the standard segmentation algorithm. Prototype test has been carried out to validate the effectiveness of the proposed method.

GF-3 is the Chinese Synthetic Aperture Radar (SAR) satellite mission with scientific and commercial applications, which will be launched in 2016. The ortho-rectification image combined with the DEM data can only be satisfied with the applications of the high resolution radar image, the data quantity with wide-breadth also put forward higher request to the automation processing. The Reverse-Range-Doppler method was used to ortho-rectified the SAR image of GF-3 satellite based on the Range-Doppler model in this paper. It not only ensures the correction precision, but also simplifies the iteration steps about DEM data, and improves the efficiency of automatic processing.

The edges of the shadow region are blurred in the SAR image due to the moving of the radar during data collection. This phenomenon becomes obvious in the High Resolution SAR images. Shadow enhancement is of great value for ATR especially when the scattering centers of the target itself are not clear. In this paper, an approach for shadow enhancement in the SAR images for targets with plat structures is presented. And experiments on the Mini-SAR data test the validity of the approach.

Two approximations, center-beam approximation and reference digital elevation model (DEM) approximation, are used in synthetic aperture radar (SAR) motion compensation procedures. They usually introduce residual motion compensation errors for airborne single-antenna SAR imaging and SAR interferometry. In this paper, we investigate the effects of residual uncompensated motion errors, which are caused by the above two approximations, on the performance of airborne along-track interferometric SAR (ATI-SAR). The residual uncompensated errors caused by center-beam approximation in the absence and in the presence of elevation errors are derived, respectively. Airborne simulation parameters are used to verify the correctness of the analysis and to show the impacts of residual uncompensated errors on the interferometric phase errors for ATI-SAR. It is shown that the interferometric phase errors caused by the center-beam approximation with an accurate DEM could be neglected, while the interferometric phase errors caused by the center-beam approximation with an inaccurate DEM cannot be neglected when the elevation errors exceed a threshold. This research provides theoretical bases for the error source analysis and signal processing of airborne ATI-SAR.

Multichannel synthetic aperture radar promises high-resolution and wide-swath imaging simultaneously. Channel error estimation is a critical step in signal processing before imaging. This letter mainly derives the Cramér-Rao lower bounds (CRLBs) for phase error estimates of three commonly used error estimators. Furthermore, to comprehensively evaluate the estimators, this letter compares both their accuracy and effectiveness. The accuracy is assessed by the maximum estimation deviation among channels, and the effectiveness is assessed by the proximity of the mean square errors (MSEs) to CRLB for phase error estimates. Finally, simulation is conducted to compare the maximum deviation as well as the MSE versus CRLB among the three estimators, under different clutter distributions and signal-to-noise ratios. Combined with the estimation accuracy and effectiveness, this letter aims to provide justifications for the proposed algorithms and gives recommendations for method selection in engineering applications.

This paper introduces a novel unsupervised estimator of equivalent number of looks (ENL) that can be applied to an arbitrary image. It avoids the assumption that homogeneous speckle will dominate the investigated image that is followed by current unsupervised ENL estimators but not always valid, especially for the complex SAR scenes with high mixture and texture. Incorporating the statistical properties of ENL data into an automatic segmentation method, we isolate the sub-class affected least by mixture and texture and suggest taking the mean value of this class as the final ENL estimate. The proposed estimator is evaluated in the experiments performed on simulated and real data from two very different sensors. It always gives better results than the other two existing methods and possesses greater adaptability.

The boundary of the shadow region is blurred in SAR images for the moving of the radar during the collection of data. This phenomenon gets particularly obvious for tall and narrow targets in High Resolution (HR) SAR images. In this work, based on the Height-Variant Phase Compensation Algorithm (HVPC) according to the property of the target like poles, an approach for shadow enhancement about tall and narrow targets in SAR images is presented. This approach uses a group of a compensation filters in azimuth direction to focus the edge of the shadow. In addition, the experiments for Terra-SAR and Mini-SAR data are implemented and the results prove the validity of the approach.

Existing low-Earth-orbit synthetic aperture radar (SAR) algorithms generally assume that the data are azimuth invariant. However, this assumption does not hold for the medium-Earth-orbit (MEO) SAR systems due to the significantly longer azimuth integration time and complex imaging geometries. As a result, the MEO SAR data cannot be processed accurately and efficiently using the existing algorithms. To solve this problem, this letter proposes a two-step azimuth perturbation (AP) method that uses the first-step AP to remove the bulk azimuth variance at the range processing stage and the second-step AP to remove the residual variance at the azimuth processing stage. As an example, an improved range Doppler algorithm with the integrated two-step AP is discussed in this letter. Simulations of an L-band MEO SAR with 5-m resolution at 10 000-km orbit height are used to demonstrate the validity and accuracy of this algorithm.

Multiple scattering may render synthetic aperture radar (SAR) image interpretation difficult, particularly when it comes to imaging of man-made structures, which can be modeled as composite scatterers. To isolate different scattering mechanism, we designed an airborne SAR experiment based on the high resolution, sub-metric to decimetric range, capabilities full-polarimetric SAR system, CARSS (Chinese Airborne Remote Sensing System), developing by IECAS. The imaging results are quite accord with the theoretical analysis. With polarimetric target decomposition, we can simply distinguish the different scattering mechanism. The idea to interpret the man-made targets as the combination of simple scattering mechanisms is supported by the experiment results.

The geometry distortion phenomenon of SAR imaging arises in the mountainous scenarios due to the presence of strong terrain slopes. Because of phase discontinuities or the absence of valid phase, for single pass interferometric SAR (InSAR), it is difficult to recover accurate digital elevation model (DEM) in such areas. Fusion of two or more different aspects of InSAR data is practicable to deal with this problem. In this paper, the processing procedures of airborne InSAR data are presented. In order to decrease the processing error of every single aspect data, an iterative motion compensation (MOCO) method is used. Besides, the interferometric phase of shadow area is linearly complemented before phase unwrapping to avoid error spreading. Experimental results using two anti-parallel aspects of airborne InSAR data validate the feasibility of fusion.

HJ-1C satellite was launched in November 2012 as the first civil spaceborne SAR system of S band in China. Since the image geolocation is essential for the quantificational study of the system specifications of HJ-1C, the paper focuses on the geolocation of HJ-1C SAR image, including the systematic position accuracy of HJ-1C satellite image, and the higher position accuracy with GCPs(ground control point). In addition, a novel equivalent-range-doppler method is presented in this paper to geolocate HJ-1C satellite image using only one GCP.

A novel parameter estimation algorithm of fast moving targets using single-antenna airborne synthetic aperture radar (SAR) databased on desampling and Radon transform (RT) is introduced in this letter. First, the dual-channel data are constructed by desampling the single-antenna airborne SAR data in the azimuth direction. Then, the clutter and the spectrum aliasing of the moving target can be cancelled by coherent subtracting. As a result, the moving target trajectory exhibits a single curve in both the range-compressed and the range-Doppler domains. Second, range cell migration correction is adopted to eliminate the range curve and parts of the range walk. Owing to the Doppler ambiguity, the moving target trajectory becomes a straight line. Third, the desampled Doppler ambiguity number and the Doppler rate of the moving target can be calculated by the slope of the line, which is measured by RT. Finally, along- and across-track velocities of the moving target are further obtained. The effectiveness of the proposed scheme is validated by the simulated and real data.

The signal processing of the medium-Earth-orbit synthetic aperture radar (SAR) is more challenging than that of the current low-Earth-orbit SAR because the imaging geometry is more complicated, and the range and azimuth variances are more severe. This paper deals with these imaging problems in three aspects. First, an advanced hyperbolic range equation (AHRE) is proposed for the first time, which is more precise for a spaceborne SAR than the conventional hyperbolic range equation (CHRE). Second, the point target spectrum based on the AHRE is analytically derived, which is useful for developing efficient SAR processing algorithms. Third, the well-known nonlinear chirp scaling (NLCS) algorithm is modified according to this new spectrum, and the so-called AHRE-based advanced NLCS (A-NLCS) algorithm is established. The simulation results validate the correctness of our method for L-band SAR systems at altitudes from 1000 to 10 000 km with an azimuth resolution around 3 m. It is also shown that the A-NLCS algorithm has better performance than the CHRE-based algorithms in longer integration time cases. Therefore, we recommend the A-NLCS algorithm for a spaceborne SAR with a lower frequency, finer resolution, and higher satellite altitude.

A synthetic aperture radar (SAR) raw data simulator is an important tool for testing the system parameters and the imaging algorithms. In this paper, a scene raw data simulator based on an inverse ω-kalgorithm for bistatic SAR of a translational invariant case is proposed. The differences between simulations of monostatic and bistatic SAR are also described. The algorithm proposed has high precision and can be used in long-baseline configuration and for single-pass interferometry. Implementation details are described, and plenty of simulation results are provided to validate the algorithm.

Processing bistatic SAR image all through numerical calculation is the concept of NuSAR. In this paper, the block scheme of NuSAR is modified to handle the translational variant case. Then a new calculation method of NuSAR is provided, which saves half of the memory compared with the existing calculation method. The provided new NuSAR is practical and can handle the spaceborne-airborne configuration.

In this paper, the effects of multi-path scattering mechanisms on SAR image is deduced through range Doppler algorithms (RDA). The conclusion that the cloud phenomenon appeared due to the multi-path scattering mechanisms is detained. Through the analysis, the cloud caused by the multi-path in the down range is corresponding to focus mechanisms and the cloud appeared azimuth is non-focus. At last, the shooting and bouncing ray (SBR) technique is employed to calculate the scattering of the cylinder cavity and by combining with range Doppler algorithms (RDA), the SAR image of a cylinder cavity with underside closed is precisely given, considering the effects of the multi-path scattering mechanisms in different azimuth.

As the Doppler history depends both on the transmit range and the receive range in bistatic SAR, those targets, who have the same bistatic range but different locations, will have different doppler history. So the unknown terrain undulation will cause defocusing in bistatic SAR imaging. This paper firstly analyzes which parameters in bistatic configuration affect the defocusing severely, and then shows the simulation results to testify the analysis. Besides, it points out the 3D imaging ability (though very weak) of bistatic SAR based on the defocusing phenomena. And Finally the 3D imaging results are also exhibited.

Forward-looking imaging has many potential applications, but it is impossible with the usual monostatic synthetic aperture radar (SAR) principle. Through the bistatic SAR configuration, forward-looking imaging can be realized for one of the bistatic platforms. This letter designs a bistatic configuration with a stationary transmitter and a forward-looking airborne receiver. It then analyzes the 2-D resolution and finds out which geometric parameter affects the imaging ability mostly. Besides, it gives out the signal formulation in the frequency domain and shows its imaging characteristics. Then, an imaging method is chosen for this special configuration, and the simulation results are exhibited, which validate the correctness of the analysis and prove the 2-D imaging ability of forward-looking bistatic SAR.

This paper first shows the 3D property of bistatic synthetic aperture radar (BiSAR) geometry, which clarifies that the algorithms for bistatic SAR should be deduced in 3D space. It then models the bistatic echo according to the 3D geometry and obtains the signal spectrum in the wavenumber domain. Based on the spectrum, the formula for the wavenumber-domain interpolation of the omega-K algorithm is deduced, and the residual phase is obtained. Then, the impacts of the residual phase, including position displacement, range, and azimuth defocusing, and a constant phase for each pixel, are explicated. Finally, the simulating results exhibited at the end of this paper validate the correctness of the analysis and the feasibility of the algorithm.

This paper deals with the imaging problem of one-stationary bistatic SAR (BiSAR) with large bistatic angle. An improved nonlinear chirp scaling (NLCS) algorithm is proposed for this BiSAR. The main work here includes three aspects. First, a range chirp scaling function for correcting the differential range cell migration correction is derived. Then, the azimuth perturbation is generated by local fit method, which makes the NLCS algorithm suitable for the large bistatic angle case. Furthermore, the negative effects introduced by the perturbation (including phase error and locality error) are discussed, and some compensation methods are proposed to enhance the capability of the algorithm. The simulating results exhibited at the end of this paper validate the correctness of the analysis and the feasibility of the algorithm.