The suppression of clutter and the positioning of underground pipelines are crucial steps in the processing of ground-penetrating radar (GPR) data. It is challenging to acquire clutter-free measured data during the radar detection process. As a result, the existing deep learning (DL) methods are primarily trained using simulation data, which limits their applicability to real-world scenarios. To address these challenges, this letter proposes an improved underground clutter suppression and pipeline positioning network. In the first stage, the model is trained using both measured data and simulation clutter-free data to enhance its ability to suppress clutter in measured data. Furthermore, in the second stage, the network is modified to accept paired, labeled simulation data, which enables more accurate pipeline positioning than the original unpaired network. Real-world data evidence demonstrates that the proposed network’s clutter suppression achieves a mean squared error (mse) of 0.006 and a peak signal-to-noise ratio (PSNR) of 34.73 dB. Additionally, the Euclidean distance error of the target clustering center coordinates is 0.82px. Compared to other methods, the performance of the proposed approach has been significantly enhanced.

Cross-modal unmanned aerial vehicle (UAV) target detection and recognition has shown great advantages under complex background, such as poor illumination, shading and deformation, etc., because cross-modal data can provide complementary information mutually. However, the real cross-modal or multi-modal data suffers from the problem of position shift and spatial deformation, that is, the image pairs are not strictly aligned, which is extremely unfriendly to downstream detection, recognition or tracking tasks. In this paper, we propose a cross-modal registration method for visible and infrared images of small UAV. Firstly, a coarse registration module based on ROI coordinates is designed to perform local registration of unregistered images, which can fully extract and utilize the small-scale feature information in the cross-modal images. Then, a semantic-based fine registration module is further proposed for fine correction and matching. In addition, we manually rearrange and relabel the Anti-UAV dataset, and experimental validations on the dataset are performed, demonstrating the effectiveness of the proposed algorithm by comparing with classical approaches.

Synthetic Aperture Radar (SAR) 3D imaging for complex structural buildings is a hot research topic in the SAR imaging field. Tomographic SAR and multi-aspect SAR are two main imaging modes for SAR 3D imaging. The former one requires repeat passes and is mainly used in spaceborne SAR. The latter one, which is discussed in this paper, observes the target area from multiple aspect angles and obtains multi-aspect target scattering features. However, it still faces some challenges for 3D imaging of complex structural buildings. The main problem is that the elevation resolution is limited by the anisotropic scattering of the target. To solve this problem, we propose a new 3D imaging method that combines the high accuracy elevation inversion capability of interferometric SAR and the elevation ambiguity resolving capability of multi-aspect SAR. Real data validates the proposed method.

Due to the rapid development of cognitive radar and the complicated electromagnetic environment, traditional anti-jamming decision-making methods are no longer suitable to modern electronic counter-countermeasures. Reinforcement learning brings a novel solution to this problem. In this paper, a method based on deep reinforcement learning is applied in the anti-jamming decision-making system of cognitive radar. We construct the environment model for cognitive radar and propose a modified deep deterministic policy gradient algorithm for decision-making. The experimental results demonstrate that the proposed method is effective in the application of anti-jamming decision-making system of cognitive radar. Furthermore, the performance analysis shows that the proposed algorithm converges faster than other classical algorithms and more suitable to high-dimensional state and action space problems.

Spaceborne spotlight SAR mode has advantages of relative wide coverage and high resolution. Moving target imaging using spaceborne SAR system is important in both civil and military applications. Current researches focus on ground and maritime target like vehicle and large ship, the topic about aeroplane is rear. Therefore, based on Chinese GF-3 spotlight SAR SLC data, a new moving aeroplane imaging method is proposed. the moving target signal model in spotlight SAR SLC is deduced, the residual range migration and azimuth quadratic phase are then analyzed. it turns out that the residual range cell migration and quadratic phase relate to the relative speed. then, the moving aeroplane can be focused iteratively via tuning the relative speed parameter. The GF-3 spotlight SAR data of a non-cooperative aeroplane is used to validate proposed method.

In this article, we present a novel synthetic aperture radar (SAR) 3-D imaging strategy using circular Interferometric SAR (InSAR) data. Our approach improves upon the Holographic SAR tomography (HoloSAR) techniques by eliminating the need for multi-baseline data collection nor residual motion error correction over a long curvilinear aperture. This may provide a simple yet effective 3-D imaging solution for perturbed airborne radar platforms. The key innovation is the first utilization of multi-aspect SAR interferograms to invert the 3-D or 4-D (3-D spatial coordinates (x, y, z) and radar azimuth angles θ) scattering power distribution of the imaged scene. Our fundamental assumption is that the imaged scene conforms approximately to a random volume scattering model. Thus, we can use the projection-slice theorem to establish a mathematical relationship between the multi-look interferograms and the scene’s 3-D/4-D scattering power distribution. The fundamental concepts and resolution theory of this new 3-D inversion strategy are developed in 3-D K-space using Fourier aperture synthesis theories. Then, we designed two algorithms for reconstructing a 3-D image: a filtered back-projection algorithm for reconstructing isotropic targets, and a compressed sensing imaging method for reconstructing large-scale targets with anisotropic behaviors. Finally, we verified the feasibility of proposed methods through experiments in real airborne scenarios.

Multi-aspect SAR can obtain more backscatter information about the target by observing targets from different azimuths by means of radar. More and more attention has been paid to the analysis of typical target characteristics in SAR scenes. The target characteristics mainly include the anisotropy and isotropy characteristics of the target. In the process of multi-aspect analysis, SAR images from different angles are routinely used to analyze typical targets in the area. In this article, the SAR images of different angles are filled into the matrix for characteristic analysis. Secondly, the full-angle data is a 360-degree coherent image of the scene. For the established matrix, we randomly remove some data to reverse the target characteristics. Calculate the feature matrix under constraint conditions by combining the matrix with the collaborative filtering algorithm, and finally get the predicted value of the missing angle. C-band circular SAR data is used to validate our method.

Due to the advantages of long-time and multi-angle observation, ground moving target detection with circular synthetic aperture radar (SAR) has recently attracted lots of interest from researchers. Our team has previously proposed the logarithm background subtraction algorithm for moving target detection in single-channel circular SAR. Its principle is that the background image (static clutter) is obtained by median filtering of the image sequence, and the foreground image (moving target) is obtained by subtraction. To further improve the performance of background and foreground separation, we introduce the low-rank sparse decomposition (LRSD) method into the previous framework. A new algorithm based on LRSD and adaptive threshold detector (ATD) is proposed in this letter. First, this letter introduces entropy metric to optimize parameters in LRSD to obtain better background and foreground separation. Second, since the statistical distribution of the foreground image is unknown, the constant false alarm rate (CFAR) detector cannot be applied to the foreground image. Therefore, an adaptive threshold detector (ATD) built on Otsu is presented in this letter, which is independent of image statistical properties. The final detection result is obtained via clustering using the modified density-based spatial clustering of applications with noise (DBSCAN) method. The effectiveness of the proposed algorithm is verified by the experimental results on the airborne X-band Gotcha Volumetric SAR Data.

In recent years, multi-angular SAR is widely researched, including wide angle SAR and circular SAR. These kinds of SAR working modes can detect the aspect dependent scattering characteristic of the target. However, most of the researches are concerned about the amplitude information. The multi-angular phase characteristic is also an important and useful information that can be obtained from the raw data or the image. In this paper, the multi-angular phase characteristic of canonical structures is analyzed by electromagnetic simulation. Then a man-made target detection method is proposed based on the multi-angular phase characteristic. The method is validated by an X-band SAR chamber data and the GOTCHA X-band circualr SAR data. The result preliminarily shows the ability of phase on analyzing the anisotropic scattering.

Multi-angular SAR can be used to obtain the information of target scattering characteristics at different aspect angles. The scattering anisotropy extraction attracts more attention but the method is not much. And because of background noise, the anisotropy extraction based on aspect entropy is not good. Low-rank matrix decomposition is widely used in the change detection process of SAR images. The most important thing is that it can distinguish the strong point target from the background image and the sparse matrix obtained by decomposition eliminates the sidelobe noise of the target. In this paper, firstly we proposed the application of low-rank matrix decomposition to multi-angular SAR images to analyze the scattering characteristics and then the coefficient of variation is used to validate and quantify the anisotropy of the target in the scene after low-rank matrix decomposition. The anisotropy quantization result is less affected by noise than aspect entropy. The Gotcha X-band circular SAR data is used to validate the method.1

Reconstructing a three-dimensional (3-D) target model from a collection of multi-aspect SAR images has been a hotspot. When imaged from different viewing aspect angles, a 3-D target will project onto different 2-D locations and present different geometric shapes. Echoes from the targets generate the non-background areas in the SAR imagery, so the geometry of the background areas can be used to restrict the possible 3-D shapes and positions of the targets. In this paper, we develop a background constraint for checking the consistency between 3-D models and the subaperture image sequence. Then a 3-D reconstruction algorithm using single-pass circular SAR (CSAR) imagery is proposed. The algorithm iteratively removes the incompatible voxels from an initial 3-D model by checking the constraint-consistency of each illuminated voxel in the model until the model converges. Also, we use a ray tracing strategy to check whether a voxel can be illuminated by the radar, so the proposed algorithm can robustly deal with the shadow effects in SAR images. The performance of the algorithm is validated using the C-band CSAR imagery acquired by the Aerospace Information Research Institute, Chinese Academy of Sciences (AIRCAS).

In conventional synthetic aperture radar (SAR) working mode, targets are assumed isotropic due to the limited aperture length. However, most of man-made targets are anisotropic. Therefore, the anisotropic scattering can help us do man-made target detection. Circular SAR (CSAR) [1] is a new SAR working mode and it can obtain the anisotropic scattering of the target by 360° observation. In this paper, the multi-angular statistical properties of targets are analyzed. The probability density functions (PDF) of the anisotropic target are various under different aspect viewing angles, while the PDFs of the isotropic target are basically stable. Then a man-made target detection method is proposed based on the multi-angular statistical property. Likelihood ratio test [2] is used to judge whether the statistical property of scattering is anisotropic or isotropic. Then anisotropic scatterings, which represent the man-made targets, can be discriminated from isotropic scatterings by thresholding. An X-band chamber circular SAR data and a C-band airborne circular SAR data are used to illustrated our idea.

Complex-valued convolutional neural network (CV-CNN) has been presented in recent years. In this letter, CV full convolutional neural network (CV-FCNN) is proposed for synthetic aperture radar (SAR) target classification, which contains only convolution layers in the hidden layer. The purpose of replacing both the pooling and fully connected layers in CV-CNN with the convolution layers is to avoid complex pooling operation and prevent overfitting, respectively. Considering the label of target is always real-valued, the magnitude of the complex vector obtained from the last convolution layer is calculated before softmax classification in the output layer. Moreover, the back-propagation formula for each layer of CV-FCNN is presented in detail. Furthermore, the complex$1\times 1$convolution layer is added into CV-FCNN to learn the cross-channel information of feature maps. The experimental results show that the average accuracy can be improved using CV-FCNN, and it is further improved using CV-FCNN with the$1\times 1$convolution layer.

The scattering of an anisotropic target is aspect dependent. Circular SAR (CSAR) can observe the scattering behavior in different aspect angles. In this paper, we propose an anisotropy scattering analysis method based on the likelihood ratio using CSAR data. CSAR data is used to provide sub-aperture images in different aspect angles. The likelihood ratio is defined as the ratio of the conditional probability under two hypotheses, anisotropic and isotropic. Anisotropic and isotropic scatterings can be discriminated by the value of the likelihood ratio. The scattering direction of the anisotropic scattering can be obtained by using our method too. We use a C-band CSAR data, which is acquired by the Institute of Electronics, Chinese Academy of Sciences (IECAS) to validate our method.

Circular Synthetic Aperture Radar(CSAR) has become a hotspot with its characteristic of elevation plane resolution and all-aspect observing ability. Digital elevation model (DEM) extraction in urban arears by using single-pass CSAR data without requiring additional knowledge is a subject of interest. The target, whose real height is not equal to the reference imaging height will project to different locations after imaging in different sub-aperture. In this paper, the quantitative relationship between offset of imaging points and height difference is deduced theoretically in the real scene, where the airborne SAR platform trajectory is not a standard circle. DEM of an area is presented using the data acquired by the Institute of Electronics, Chinese Academy of Sciences (IECAS). Compared with the DEM provided by the German Aerospace Center (DLR) with 1m absolute height error, the effectiveness of the proposed method is verified.

Circular synthetic aperture radar (CSAR) can provide distinctive multi-aspect anisotropic scattering signatures. However, it is impossible to retain the anisotropic signatures in a SAR image that combines all the subapertures coherently or incoherently. In this letter, we propose a polarimetric CSAR anisotropic scattering detection framework to characterize multi-aspect and fully polarimetric SAR signatures of point-like and distributed targets. We applied this framework to quantify and rank media polarimetric scattering dissimilarity over all aspects and to determine whether the most different one shows anisotropy by use of constant false alarm rate (CFAR) detection. Furthermore, we demonstrated the monotonicity of CFAR detection function and incorporated this function to decrease the complexity of the anisotropic scattering test. Our algorithm was validated and applied to a set of airborne P-band fully polarimetric circular SAR data acquired by the Institute of Electronics, Chinese Academy of Science (IECAS). The results indicate the framework can retain anisotropic scattering and extract a series of new multiaspect polarimetric SAR signatures for terrain classification.

Circular synthetic aperture radar (CSAR) can provide distinctive multiaspect anisotropic scattering signatures. However, it is impossible to retain the anisotropic signatures in an SAR image that combines all the subapertures coherently or incoherently. In this letter, we propose a polarimetric CSAR anisotropic scattering detection framework to characterize multiaspect and fully polarimetric SAR signatures of pointlike and distributed targets. We applied this framework to quantify and rank media polarimetric scattering dissimilarity over all aspects and to determine whether the most different one shows anisotropy by the use of constant false-alarm rate (CFAR) detection. Furthermore, we demonstrated the monotonicity of CFAR detection function and incorporated this function to decrease the complexity of the anisotropic scattering test. Our algorithm was validated and applied to a set of airborne P-band fully polarimetric circular SAR data acquired by the Institute of Electronics, Chinese Academy of Science. The results indicate that the framework can retain anisotropic scattering and extract a series of new multiaspect polarimetric SAR signatures for terrain classification.

A new ground moving target indication algorithm for single channel Circular Synthetic Aperture Radar (CSAR) is presented and evaluated by airborne CSAR dataset. The algorithm is based on overlap subaperture magnitude images. The subaperture image can be regarded as the background image (clutter) plus the foreground image (moving target). The background image is obtained by median filter. Then the moving targets can be detected by using the subaperture magnitude images subtracting the background image. The algorithm is tested by GOTCHA GMTI dataset. The algorithm is capable of real-time processing and detecting multiple moving targets. And the initial results is presented in the paper.

This paper considers techniques of aspect feature extraction and application of targets taken over one or more wide-angle apertures. Compared with traditional narrow-aperture Synthetic Aperture Radar (SAR), multi-aspect SAR can provide observation of targets in more azimuth directions. Radar backscattering is typically characterized by location and azimuth directions. Thus target anisotropic scattering properties can be exploited with multi-aspect SAR. Firstly, data inversion method is proposed to extract the Radar Cross Section (RCS) of target from each azimuth direction independently. And then, target aspect features which are target scattering persistence angle, peak value and main scattering direction are defined. These parameters' availability is validated by experimental data which is collected in chamber and airborne data. Finally, adaptive imaging method based on target scattering persistence angle is proposed to improve the existing Circular-SAR (CSAR) imaging methods. Additionally, porlarimetry is combined with target aspect features to realize the power lines extraction. Above applications of target feature extraction is achieved with airborne data.

In this letter, we propose an imaging algorithm for the holographic synthetic aperture radar tomography in the circumstance of sparse and nonuniform elevation circular passes. Considering the anisotropic behavior of scatterers and the off-grid effect of sparse signal recovery, the algorithm combines the 2-D adaptive imaging method for circular SAR and the sparse Bayesian inference-based method for elevation reconstruction. For each circular pass, the azimuth-range 2-D image can be formed by the adaptive imaging method, which depends on the preretrieved maximum azimuth response angle and the azimuth persistence width. To deal with the off-grid effect in elevation reconstruction, which is caused by the deviation between the true scatterers and the discretized imaging grids, the off-grid sparse Bayesian inference method jointly estimates the scatterers and elevation off-grid error by applying their hierarchical priors. Compared with the conventional compressive sensing method that does not concern the off-grid effect, the proposed algorithm can provide more accurate 3-D reconstruction for pointlike targets, which is verified by the real-data experiments.

Downward looking sparse linear array three-dimensional synthetic aperture radar (DLSLA 3-D SAR) can obtain 3-D scene properties and has broad application prospects. However, the reconstruction of cross-track dimension usually suffers from incomplete observation, which is caused by the non-uniformly and sparsely distributed virtual antenna phase centers. By formulating the cross-track reconstruction into the problem of sparse signal recovery, we introduce two kinds of gridless sparse recovery (GL-SR) methods to DLSLA 3-D SAR cross-track imaging, i.e., atomic norm minimization (ANM) and gridless SPICE (GLS). Compared with the conventional grid-based sparse recovery (GB-SR) methods, which assume that the scatterers are exactly on the discretized grids, the GL-SR methods can avoid the off-grid effect. Experiments compare the performance of GB-SR and GL-SR methods for DLSLA 3-D SAR cross-track reconstruction.

Multi-aspect PolSAR data contains polarimetric properties from different look angle. Multi-aspect polarimetric information can be applied in geometric measurement, target identifying, precise classification. In order to characterize anisotropic target, anisotropic and isotropic scattering need to be separated from the raw data. A detecting-removing-incoherent-adding (DRIA) framework, presented in Li Yang's doctoral dissertation, suggests to remove the anisotropic scattering, gain a removal series and incoherent integrate the reserved data. In this paper, in order to identify anisotropic target, an anisotropic scattering model is raised. An improved detection and feature retrieval method is presented base on DRIA framework. The equivalent number of looks (ENL) used in Li Yang's dissertation is proved to bring measurement error to the result. The anisotropic scattering can be correctly identified after the error is restored. Two kinds of maximum-likelihood ratio are proved to gain the same result in sort. Three features are retrieved from the removal series to describe the anisotropic scattering. The experimental data is circular SAR (CSAR) data acquired by the Institute of Electronics airborne CSAR system at P-band.

The target feature is sensitive to the aspect angle of SAR observation, making the interpretation and target recognition of the SAR image difficult. The information acquired from a certain aspect angle is partial and incomplete, and the multi-aspect observations have the potential to improve the SAR performance in this aspect. Three topics of fine feature description of multi-aspect SAR observations are discussed, and they are the 3D information extraction, the optimum imaging strategy for anisotropic scatterers, and the multi-aspect scattering feature extraction. The initial results of the real P band airborne circular SAR (CSAR) data and the turn table data show that multi-aspect SAR observations have the encouraging potential capability in target fine feature description.

Multi-aspect observation can obtain the target's multi-aspect scattering feature, and has potential to improve the SAR performance in the applications of target classification and recognition. This paper presents a time-division scanning mode based on Single Phase Centre Multiple Azimuth Beam (SPC-MAB) technique. By applying this new mode, Synthetic aperture radar (SAR) can obtain multi-aspect data from continuous scenarios through linear flight. We take six channels as an example, and analyzing the performance of the time-division scanning mode and two conventional SPC-MAB modes. According to noise equivalent sigma zero (NESZ) and azimuth ambiguities to signal ratio (AASR) simulation results, the time-division scanning mode shows better performance than conventional modes. The time-division scanning mode will be applied to the multi-aspect data acquisition vehicle-mounted system.

Compared with traditional linear Synthetic Aperture Radar (SAR), Circular-SAR(CSAR) can achieve 360-degree observation of targets. For this reason, CSAR is the best mode of SAR to exploit target feature extraction on azimuth directions. Over larger azimuth angular extents the energy reflected by targets is, in general, not uniform and most targets exhibit only limited scattering persistence. The algorithm proposed in this paper extracts the width of target persistence angle which is defined as target azimuth-angle sensitivity. It is an effective parameter to distinguish different types of anisotropic targets. And it is independent of azimuth orientations for its measurement of width of target persistence angle. This feature is extracted by data inversion in this paper. Compared with the sub-aperture approach which can also extract the anisotropic scattering properties of target, data inversion approach in our algorithm can provide more accurate results and finer curve. This target multi-angle scattering sensitivity feature extraction algorithm is validated by experimental data which is collected in chamber.

Downward-looking sparse-linear-array 3-D synthetic aperture radar (DLSLA 3-D SAR) cross-track reconstruction usually suffers from incomplete observation and limited resolution. The incomplete observation is caused by the sparse and nonuniform distribution of the equivalent antenna phase centers (APCs) due to the array elements' installation location restriction, loss, or deviation. Sparse recovery methods provide a solution with improved resolution from the incomplete observation for the 3-D imaging scene that behaves with spatial sparsity. However, conventional grid-based sparse recovery (GB-SR) methods are under the assumption that the scatterers are located on the discretized grids; otherwise, the off-grid effect or basis mismatch problem will occur. In this letter, we propose a reweighted scheme-based gridless sparse recovery (GL-SR) method, i.e., reweighted gridless sparse iterative covariance-based estimation (RGLS), for DLSLA 3-D SAR cross-track imaging. The proposed method possesses the merits of gridless SPICE (GLS), i.e., free of off-grid effect and user parameters, and has a statistically more appealing property than GLS by adopting the reweighted scheme. As seen from the experiments that compare the performance of GB-SR and GL-SR methods for DLSLA 3-D SAR cross-track reconstruction, the proposed method performs outstandingly under the circumstance of sparse and nonuniform APCs' distribution.

In circular synthetic aperture radar (CSAR) mode, a 2-D image contains 3-D spatial information about the imaged targets. This letter describes an attempt to estimate the height profiles of power lines with slightly varying heights based on 2-D CSAR imagery. First, according to the characteristics of a 2-D CSAR image of a power line with constant height, an approximate formula is deduced to calculate the height of the power line. This formula can also be used to calculate the height of several power lines with the same constant height. Second, a three-step processing method is proposed for height profile estimation of power lines with slightly varying heights. The first step is to extract regions of power lines automatically. The second step is to obtain the initial heights of the power lines according to the deduced approximate formula. The third step is to estimate the final height profiles of the power lines. Finally, experimental results have shown that 2-D imaging results of power lines with slightly varying heights are well focused when the height profiles estimated by the proposed three-step method are used.

The first airborne Circular SAR data acquisition experiment of China was carried out in Sichuan by the National Key Laboratory of Science and Technology on Microwave Imaging (MITL), China. Data was acquired using the MITL's P-band, fully polarimetric SAR system along a circular trajectory with the beam spotted on the same area. Compared with the conventional SAR along a straight path, this imaging mode mainly has the following attractive features. First, observing from all directions can help for a better understanding of the scattering properties of targets. Second, the wide angular aperture obtained via flight in a circular track makes possible high resolutions with low frequency band. Third, the aspect angle diversity inherent to the circular track allows for a 3-D target reconstruction. This paper presents the SAR processing of such data, and shows several results to analyze the potentials and limitations of such imaging geometry.

This letter presents a new algorithm for squint circular synthetic aperture radar (SAR) (CSAR) imaging, which is an extension of the well-known range migration algorithm. Due to the circular trajectory, the spatial frequency domain data of squint CSAR cannot be readily obtained via fast Fourier transform, as conventional SAR with straight path does. This method first employs along-track varying system kernels and filters to transform the raw data to the polar spatial frequency domain. Then, it uses an interpolation algorithm to convert the polar samples into rectilinear samples. Implementation aspects, including sampling criteria, resolutions, and computational complexity, are also assessed in this letter. The proposed algorithm is validated both numerically and experimentally.

The aperture of 360° gives circular synthetic aperture radar (SAR) (CSAR) the capability to detect hidden target when its orientation is unknown. Subwavelength resolution can also be achieved when the target in the spotted area is observed under a complete circular aperture. Furthermore, the aspect angle diversity inherent to the circular trajectory makes possible a 3-D target reconstruction. However, the latter two potentials require certain target reflectivity homogeneity. For a highly directive scatterer, it has no resolving ability in the direction normal to the data collection plane. In this letter, a new interferometric CSAR method is presented to enhance the tomographic imaging capability for highly directive scatterers without sacrificing other scatterers' resolutions. This method takes advantage of the coherence and the phase difference between a pair of 3-D SAR images formed from data collected at two separate circular apertures to eliminate targets that focused at a wrong elevation. In addition, it uses two different transmit frequencies to solve the problem of phase cycle ambiguities. Finally, simulation results validate this new approach.