Image classification has surpassed human performance in numerous domains. However, in real-world scenarios, challenges such as storage constraints, privacy concerns, and commercial data protection necessitate the enhancement of existing models to accommodate new class classifications — a critical area of research known as class incremental learning. Data replay stands out as a pivotal technique within this domain. The DeepInversion algorithm ingeniously employs classification networks to generate training set images of notable quality. In this study, we improve the DeepInversion algorithm, leveraging a pre-trained model to yield superior quality playback images for class incremental learning. Within the context of class incremental learning, we introduce a cosine orthogonal classification loss function, formulated on the basis of linear layer analysis, to guide image generation and augment class incremental learning. This loss function is designed to ensure the mutual orthogonality of all class centers while minimizing intra-class distances. In the realm of knowledge distillation, we harness a combination of limited real images, a profusion of synthesized images, and Mix virtual images to facilitate feature cosine distance distillation. Sufficient comparative experiments and analyses with similar latest methods underscore the efficacy of our proposed approach, and obtain the SOTA results. The code for the paper can be found at http://github.com/YunXiaooooo/Class-incremental-learning-based-on-playback-images-generated-by-classification-network

This letter proposes a coherent integration approach for weak target detection with constant radial jerk in a low signal-to-noise (SNR) environment. In the proposed approach, the second-order Keystone transform (SoKT) is first employed to correct the quadratic range migration (QRM) induced by the radial acceleration of a target. Next, a velocity-jerk matched filter is constructed to eliminate the residual target radial velocity and jerk terms. Finally, the coherent integration is achieved through non-uniform fast Fourier transform (NuFFT) along the slow time. Compared with the conventional methods, the proposed approach can obtain comparable accumulation performance with a lower computational complexity. Simulation and real-measurement experiments are conducted to demonstrate the effectiveness and reliability of the proposed approach in low SNR environments.

As a wideband radar system, spaceborne synthetic aperture radar (SAR) has been widely applied in multiple applications, such as maritime surveillance and terrain observation. However, with the increase of electromagnetic devices, spaceborne SAR suffers from radio frequency interference (RFI) frequently. Many previous methods have been effective in interference mitigation, among which semiparametric methods demonstrate excellent performance and high efficiency. However, as a classic low-rank recovery method, robust principal component analysis (RPCA) usually suffers from the over-penalization problem of large singular values. Although some useful schemes were proposed to address this issue, their performance may still degrade if the low-rank characteristics of interference are not prominent. To overcome these obstacles, we first investigate the characteristics and distributions of different SAR signals and leverage kurtosis to differentiate interference and real echoes. Herein, interference tends to have a low kurtosis while the real echoes tend to have a high kurtosis. Then, we improve the low-rank recovery model with kurtosis and propose the kurtosis-based reweighted nuclear norm (KRNN) model to precisely extract interference components. Then, we derive the closed-form solution of the KRNN model via the alternating direction method of multipliers (ADMM) framework. Through the proposed KRNN method, we can effectively mitigate interference, preserve real echoes, and solve the problems of over-penalization and nonideal low-rank property. Finally, we conduct numerical experiments using the measured Sentinel-1 and LT-1 data to demonstrate the effectiveness and robustness of our proposed method.

Radio Frequency Interference (RFI) would lead to degradation of image quality for synthetic aperture radar (SAR) systems, resulting in a waste of observation resources. RFI source localization provides crucial prior information for RFI mitigation and spectrum management, and traditional RFI localization methods for multichannel SAR system suffer from localization ambiguity. This paper derives the RFI model for dual-channel SAR and investigates the mechanisms underlying localization ambiguity. A localization framework is proposed based on the varying characteristics of RFI with platform motion and the slant range difference model between dual-channels. This approach achieves localization by implementing mutual constraints among alternative optimization models. Moreover, an RFI filtering scheme is developed to facilitate the mutual cancellation of RFI between the two channels. The performance of the proposed method is validated using simulated RFI and real-measured RFI scenarios of Chinese Lutan-1 mission. The results indicate that the proposed method effectively mitigates the localization ambiguity, demonstrating high-precision localization performance. Moreover, it can reduce echo amplitude distortion and preserve degrees of freedom while removing RFI effectively.

LuTan-1 (LT-1) is the first civil L-band synthetic aperture radar (SAR) satellite constellation and comprises two identical satellites. LT-1 is designed to fulfill two main requirements, one of which the main tasks is to provide digital surface model (DSM) products covering the areas that are not available using optical satellites. The second task is to provide deformation products to support the geohazard monitoring task. For the first task, LT-1 provides a novel noninterrupted imaging technology to facilitate phase synchronization. For the second task, the orbit maintenance is implemented using a newly proposed in-plane offset semimajor axis and out-of-plane control triggered strategy. Besides, we provide the system performance analysis for the two main tasks. Finally, the first results are produced with root-mean-square-error (RMSE) of the DSM less than 0.7 m in the flat region and 6.7 m in the mountainous region. The RMSE of the first deformation products is less than 2.7 mm in the test region. The results indicate a favorable potential for terrain mapping and deformation monitoring applications.

Range ambiguity is a common issue for spaceborne synthetic aperture radar (SAR) systems, leading to image quality degradation and subsequent interpretation accuracy. Existing range ambiguity suppression methods mainly rely on raw echo domain processing with specific requirements on prior knowledge. However, most users can only obtain single-look-complex(SLC) products instead of raw echo products. Raw echo obtained from SLC through an inverse focusing process will waste additional time and space resources. This paper proposes a novel scheme to detect and suppress range ambiguity in SLC products to deal with this deficiency. The proposed method designs a detector to extract and suppress the focused range ambiguity based on the characteristic difference between the desired signal and range ambiguity in the fractional transform domain of the SLC image. Experimental results on real measured L-band spaceborne interferometry SAR system verify the detection performance of the proposed method, which is beneficial for subsequent land monitoring applications.

A spaceborne bistatic surveillance radar system, which may increase the target radar cross section through a large bistatic angle, can effectively improve the performance of small air moving target indication. However, due to the separation of the transmitter and the receiver, the clutter exhibits severe range dependence in both Doppler and spatial-angle domains. This may significantly decrease the number of independent identically distributed samples and, thus, cause the degradation of clutter suppression performance. In addition, when an air moving target flies above the sea, the broadened spectrum of sea clutter caused by the random intrinsic motion will further influence the clutter suppression performance. To deal with these problems, this article establishes a multichannel sea clutter model for a spaceborne bistatic surveillance radar system. The calculation of the positions of the iso-range contours is analyzed first. Then, the bistatic sea clutter model is established based on the directional wave spectrum, the monostatic–bistatic equivalence theorem, and the two-scale scattering model, where the Earth rotation and the range ambiguity of clutter are considered. Finally, the range dependence and the spectrum characteristics of bistatic sea clutter are discussed, and the clutter suppression performances under different azimuth bistatic angles, different elevation angles, and different sea states are analyzed. In addition, the influence of clutter range ambiguity is analyzed based on the simulations.

As one of the main statistical characteristics of clutter, the clutter amplitude distribution plays an important role in the accurate modeling of spaceborne multi-channel radar signal as well as the subsequent, maritime radar target detection. In this paper, considering that the space-time adaptive processing (STAP) technology is usually applied to accomplish the main-lobe clutter rejection in a space-borne surveillance radar, the influence of different clutter amplitude distributions on STAP in a spaceborne multichannel radar system are analyzed. Firstly, a spaceborne multi-channel clutter model is established based on radar equation and clutter space-time steering vector. Then, Rayleigh distribution, Weibull distribution, lognormal distribution, K distribution, generalized Pareto distribution, and IG-CG distribution are used to fit the clutter amplitude. Finally, the effects of these clutter amplitude distributions on STAP performance are analyzed, respectively. The simulation results show that in the case of the same clutter power, the influence of different clutter amplitude distributions on STAP performance is approximately the same.

Multiple-input–multiple-output (MIMO) technique has recently received great attention due to its promising prospect for ground moving target indication (GMTI) applications in a synthetic aperture radar (SAR) system. In this article, an SAR-GMTI processing technique in an MIMO-SAR system with dual-channel configuration is proposed. In the proposed method, up and down chirps are adopted as the transmitted waveforms. Then, the baseline estimation and clutter rejection are simultaneously accomplished in the raw Doppler data domain, after which two separated range-compressed echo signals corresponding to two transmitters can be obtained without the influence of cross-correlated interferences of the clutters. Subsequently, a ground moving target can be detected, clustered, finely refocused based on the Doppler chirp parameter estimation, and accurately relocated by using the along-track interferometry (ATI) processing technique. Simulated experiments are implemented to validate and evaluate the feasibility of the proposed method.

For a spaceborne or an airborne synthetic aperture radar (SAR) system, it may be easily interfered with radio frequency interferences (RFIs) due to the increase demand for frequency occupation. The RFIs introduce artifacts in the focused SAR imagery, resulting in high noise floor and erroneous interpretations. Previous works on narrowband RFI mitigation mainly focused on single-channel SAR systems, while with the development of SAR imaging technology, the high-resolution wide-swath (HRWS) imagery technology now reaches its maturity to finally take shape in current SAR systems. To obtain HRWS images, the multi-channel SAR (MC-SAR) system has been employed to tackle the contradictory requirements for both high resolution and low pulse repetition frequency (PRF). In this paper, we propose a novel interference-mitigation model and analyze the low-rank property of the narrowband RFI for HRWS SAR systems. Then, we employ an image-domain sparse regularization to protect the real echoes of the SAR system and mitigate the RFIs by solving the low-rank recovery problems of RFIs. The real airborne SAR data is used to demonstrate the effectiveness of the proposed method.

In this letter, a novel method is proposed to suppress the terrain fluctuation clutter based on adaptive broadening processing. In the proposed algorithm, the flat interference phase is first compensated according to the priori radar system parameters. Then, according to the space-time trajectory distribution of clutter edge, the level of clutter Doppler spread in virtue of crab effect is estimated by a cost function related to the clutter eigenvector matrix. Finally, after calculating the clutter suppression weight vector according to the broadened clutter subspace, the non-stationary ground clutter can be robustly rejected. The validity of the proposed method is verified by both the simulated and real-measured multichannel radar data.

The spaceborne high-resolution wide-swath synthetic aperture radar (HRWS-SAR) system generally does not meet the optimal SAR imaging configuration, and thus, it is necessary to apply digital beam-forming filtering technology to restore the nonuniformly sampled signal into the uniform grids. However, in practice, because of the influences of temperature, receiving machine, and other error factors, the channel errors may possibly exist, causing the SAR image to be smeared. To address this issue, this letter proposes a novel algorithm based on the hybrid artificial fish school algorithm–glowworm swarm optimization–genetic algorithm (AFSA-GSO-GA) to address the channel imbalance issue. First, coarse HRWS-SAR imaging processing is performed to obtain the positions of the Doppler ambiguity components. Then, according to the designed cost function, the hybrid AFSA-GSO-GA algorithm is used to realize the channel phase error estimation. Finally, a well-focused SAR image can be obtained after channel balance. The effectiveness of the proposed method is validated by both simulated and real SAR data.

Due to the limited size, carrying capacity, power-aperture product, and high hardware cost of satellite platform, the traditional single-platform spaceborne radar system encounters the problems of poor target minimum detectable velocity (MDV) performance, considerably deteriorating the moving target detection performance. To improve the air moving target indication (AMTI) performance, especially for a weak target, distributed space-based radar system (DSBR) becomes a good candidate due to the longer along-track baseline (ATB) and spatial power synthesis. However, due to the sparse configuration of radar baseline distribution, the detection performance of air moving targets (AMTs) will be restricted by many practical factors in an actual DSBR system. In this paper, multi-channel signal models of an observed moving target and ground clutter are accurately established in a DSBR framework, where the error influences of cross-track baseline (CTB), terrain fluctuation, and channel inconsistency response are considered. Then, the influence of the non-ideal factors, including the channel noise, long-intersatellite ATB, long-intersatellite CTB, synchronization errors, and interchannel amplitude and phase inconsistency errors, on the AMTI performance is analyzed term by term. The simulation results provide the useful guidance for the system design of a DSBR with the AMTI tasks.

Synthetic aperture radar (SAR), as a wideband radar system, may be subject to strong interferences with a variety of signals. The narrowband interference (NBI) exemplifies the most typical of its kind, such as the form of radio frequency interference (RFI). With the development of SAR imaging technology, the high-resolution wide-swath (HRWS) imagery technology now reaches its maturity to finally take shape in current SAR systems. To obtain HRWS images, the multichannel SAR (MC-SAR) system has been employed to tackle the contradictory requirements for both high resolution and low pulse repetition frequency (PRF). Previous interference methods focused on single-channel SAR systems and few research works for MC-SAR systems. In this article, we first derive a new interference-mitigation model for HRWS SAR systems and conclude that the low-rank property of the NBI is suitable for MC-SAR systems. Then we employ an image-domain sparse regularization to protect the real echoes of the SAR system and mitigate the NBIs by solving the low-rank recovery problems of NBIs. Also, the MC-SAR system errors are further taken into account as a measure for our method’s practical applicability. Finally, the real SAR data is used to demonstrate the effectiveness of the proposed method.

Radio frequency interference (RFI) is a significant threat to accurate microwave remote sensing. The RFI signals manifest themselves in unpredictable locations and patterns in the image, which will cause measurement distortion, image degradation, or even lead to wrong retrievals of the geophysical parameters. Accurate detection of RFI artifacts is a prerequisite step to preserve the overall quality of remote sensing quality. In this paper, a semantic cognitive enhancement network for RFI signature detection is proposed. It employs an encoder-decoder architecture, which incorporates the atrous spatial pyramid pooling, Depthwise convolution, and self-attentional mechanism. Rather than detecting the existence of RFI artifacts for an entire image, the proposed scheme can realize RFI recognition in a pixel-wise manner without setting predefined thresholds. Extensive experimental results on diverse scenarios in Sentinel-1 images with various RFI types are provided, which demonstrates robust detection performance for both strong and weak interference without requiring a large number of training samples.

In this article, an effective clutter suppression algorithm is presented for the space-borne azimuth multiantenna high-resolution and wide-swath (HRWS) synthetic aperture radar (SAR) system, which is based on the null-space technique. First, the echo of bistatic geosynchronous-low earth orbit (GEO-LEO) azimuth multiantenna HRWS SAR-ground moving target indication (GMTI) system is utilized to deduce the coarse-focused image of moving targets and clutter, where the Chirp Fourier transform (CFT) in azimuth is involved. Then, the matrix form can be utilized to describe the coarse-focused image of the multiantenna SAR system and the corresponding covariance matrix can be estimated. After that, the null-space is constructed using the covariance matrix corresponding to clutter, where at least a redundant channel freedom is required. Since the null-space vector is orthogonal to signal-space vector, it can be used to suppress the clutter. As an equivalent phase is brought by the slant velocity, the moving targets’ echo can be preserved during clutter suppression. Then, the optimization and suboptimization vectors for clutter suppression are introduced. It is worth noting that the proposed clutter suppression algorithm is robust for the antenna mismatch, that is, the antenna mismatch in phase and the corresponding position error. Finally, the theoretical investigations are validated using some simulation experiments, where the experiments for bistatic GEO-LEO and single-platform azimuth multiantenna HRWS SAR-GMTI system are both included. In addition, the real measured single-platform azimuth multiantenna HRWS SAR data experiments are also performed.

In this letter, a novel method is proposed to estimate the along-track baseline for a multichannel high-resolution and wide-swath (HRSW) synthetic aperture radar (SAR) system. First, a spatial correlation function is constructed to remove the range cell migration and Doppler broadening of ground static targets. Then, the iterative adaptive approach (IAA) is applied to iteratively estimate the along-track baseline with high precision. Finally, a high-resolution SAR image can be obtained based on the estimated baseline. The effectiveness of the proposed algorithm is validated by both simulated and real SAR data.

Spaceborne bistatic synthetic aperture radar (SAR) system with an inclined geosynchronous (GEO) illuminator and a low-earth-orbit (LEO) receiver is capable of providing a vast area of surveillance and fine spatial resolution, which presents huge potentials for future earth observation. In this letter, inclined GEO-LEO azimuth multichannel SAR (MC-SAR) system for high-resolution wide-swath imaging is investigated. Starting from modeling geometry of inclined GEO-LEO bistatic MC-SAR, the signal model is analyzed in detail for the first time, and the equivalent positions of the received channels are obtained. Then, we found the spatial-variant residual phase error cannot be compensated accurately by conventional effective phase center (EPC) processing. Moreover, because of the complex bistatic configuration, the frequency-domain imaging algorithms become extremely difficult to achieve a well-focused and phase-preserved image. Meanwhile, the time-domain back-projection algorithm (BPA) faces with a technical challenge due to the nonuniform sampling in azimuth. To address these issues, a bistatic weighted BPA (BWBPA) for GEO-LEO MC-SAR is derived and presented. Without the procedure of EPC, the BWBPA can suppress azimuth ambiguities effectively and yield phase-preserved SAR images. Simulated data results show the validity of the presented method.

Phase unwrapping, a process of recovering the real phase from ambiguous one, is one of key techniques for interferometric synthetic aperture radar (InSAR) data processing. The larger interferograms make the phase unwrapping to be a more challenging task. A novel fast phase unwrapping method is proposed in this paper which is very effective in dealing with the large scale datasets. Several sub-networks consisting of residues are constructed in the method. Therefore, smaller networks make the execution more efficient using minimum cost flows(MCF) algorithm. In addition, the method takes an effective strategy to balance all the sub-networks in order to ensure a feasible solution to each sub-network. Experiment with real data demonstrates the effectiveness of the proposed method.

Polarimetric synthetic aperture radar (SAR) Interferometry (PolInSAR) becomes a hot topic in SAR application areas, such as terrain classification. However, the system parameter design is mostly based on traditional Polarimetric SAR (PolSAR) or interferometric SAR (InSAR) methods. The accuracy of polarimetric calibration is the major factor for the PolInSAR system design, so it is necessary to analyze the affect of polarimatric calibration errors on terrain classification accuracy. In this paper, error transfer link of typical classification algorithm is investigated, then, the minimum requirements on polarimetric distortions including channel imbalance and crosstalk are investigated for terrain classification application for the first time. The result is verified by E-SAR datasets, which can provide the important theoretical basis and reference for quantitative calibration requirements of classification accuracy.

When combined with digital beam-forming (DBF) techniques, multichannel synthetic aperture radar (SAR) systems can achieve high-resolution and wide-swath SAR imaging. However, inevitable channel biases will degrade the performance of DBF in practice. To address this problem, a novel channel phase bias estimation algorithm is proposed in this letter. Theoretical analysis reveals that the signal of the first channel, which is considered as the reference channel, can be reconstructed from the signals of other channels, regardless of noise and signals outside the Doppler bandwidth. In the presence of phase biases, there is a reconstruction error after the cancellation by subtracting this reconstructed signal from the original signal. However, by minimizing the reconstruction error, the channel phase biases can be precisely estimated. The effectiveness of the proposed algorithm is validated by the experimental results.

Ship classification is an important step in maritime surveillance utilizing synthetic aperture radar images. In this paper, we focus on the classifier architecture. The paper investigates three individual classifiers, i.e., the K nearest neighbor classifier, the Bayes classifier, and the back-propagation neural network classifier from the viewpoint of discrimination measurements firstly. Then, we propose a SVM combination strategy to fuse the results of individual classifiers. Extensive experiments conducted on the TerraSAR-X SAR images validate the effectiveness of the proposed method.

When bistatic synthetic aperture radar (SAR) works in alternating bistatic/ping-pong mode, it can perform phase referencing without a synchronization link between different satellites. The key to this is to extract a compensation phase from the bistatic echoes, including oscillator phase differences. An echo-domain phase synchronization algorithm using the correlation of bistatic echoes is proposed. Because of its larger equivalent synchronization frequency, the echo-domain algorithm has higher phase synchronization accuracy than the algorithm in the image domain. Experiments using simulated bistatic echoes with real single-stripmap SAR echo validate the proposed echo-domain phase synchronization algorithm.

A radar imaging algorithm named reciprocal spectrum algorithm (RSA) is proposed in this paper to get higher resolution in azimuth direction with frequency sampling waveform. Theoretical analysis and simulation results show that the algorithm can give better performance than the tradition range Doppler algorithm (RDA) at the cost of peak value reduction at the object points in radar image.

A progressive SAR imaging technique is proposed as a novel SAR raw data processing scheme in this paper. Different from the classic SAR imaging algorithms which focus the SAR raw data as a whole regardless of the backscattering signal, the novel scheme discriminate the backscattering signal and choose the proper ones to be focused progressively according to some application context while the others will be discarded. The new scheme makes the SAR imaging algorithm not only focus the energy back to the scattering points but also form an image more suitable for some special applications. In this paper, a realization of the scheme present via Atomic Decomposition (AD)[1]–[3]. AD helps estimate the parameters of backscattered signal of a scattering point and detach it from the raw data. With the parameters, the backscattering-signal can be reconstructed and focused. Targets or scatter-points in the scene will be imaged progressively in a sequence of their energy due to the greedy natural of matching pursuit employed during AD. Therefore, the contents in the final SAR image can be controlled by an energy threshold. Besides the interested targets can be extracted easily from the background clutter and this may be helpful for the SAR image understanding