Data augmentation is one of the most straightforward methods to alleviate the problem of data scarcity in deep learning-based synthetic aperture radar (SAR) automatic target recognition (ATR). For example, novel view generation based on physical electromagnetic models or deep generative models. Besides the data scale, the quality of generated SAR images also has a significant impact on the model's performance. Therefore, it is crucial to objectively and effectively assess the quality of generated SAR images. In this work, a relative recognition accuracy score (RRS)-based evaluation criterion is proposed. The core idea of the proposed evaluation criterion is to control variables. First, the measured SAR images are gradually replaced by the generated SAR images at a progressive hybrid rate (HR) to obtain hybrid training datasets with varying HR levels. Then, the recognition model (e.g., AconvNet) is trained on both the all-measured dataset and hybrid datasets with varying HR to plot the reference recognition accuracy (RRA) and the hybrid recognition curve (HRC). Finally, the integral difference between RRA and HRC is defined as the RRS of the generated SAR dataset. Specifically, we employ proportional random sampling and multiple sampling averaging to alleviate selection bias. Experiments on the moving and stationary target acquisition and recognition (MSTAR) dataset validate the rationality and effectiveness of the proposed RRS-based evaluation criterion.

Despite the in-depth understanding of the synthetic aperture-radar (SAR) speckle and its characteristics, despeckling remains an open issue far from being solved. Deep-learning methods with supervised training have made great progress. However, reliable reference images are inconveniently accessible or even non-existent. In this paper, we propose an end-to-end self-supervised method named Speckle2Self for SAR image despeckling, which learns mapping from noisy input to clean output using only the input noisy image itself for training. We formulate the image despeckling as a masked pixel-estimation problem, where a set of masks is carefully designed. The masked pixel values are predicted by the queries of complementary masks indicating the positions of masked pixels through an attention mechanism. Transformer architecture is employed as the network backbone. In addition, a novel loss function is also derived based on the statistics of SAR images, and meanwhile, image downsampling is used to provide guarantees on the white noise assumption involved in our Speckle2Self. We compare the proposed Speckle2Self with reference methods on both synthetic and real images. Experimental results demonstrate that the proposed Speckle2Self achieves comparable despeckling performance with supervised methods, suppressing noise while maintaining structural details. Even compared with self-supervised methods, the proposed Speckle2Self still has significant advantages in SAR image-despeckling metrics.

This paper proposes a synthetic aperture radar (SAR) ship detection method based on wavelet-domain deformable convolution (WDC) and multi-head attention, built upon the Sparse R-CNN framework. First, a wavelet-domain convolution module is introduced to enhance the modeling of ship targets with diverse scales and shapes while incorporating frequency-domain information. Deformable convolution adaptively adjusts sampling locations, overcoming the limitations of traditional convolution in capturing target edges and blurred boundaries. Next, a position encoding module is employed to normalize candidate bounding box coordinates and integrate them into region-of-interest features. By providing spatial context, position encoding strengthens spatial perception and enables the subsequent multi-head attention mechanism to more effectively capture associations between targets and candidate regions, thereby improving localization accuracy under arbitrary spatial distributions. Furthermore, the original dynamic head is replaced with a multi-head attention mechanism. Through position-encoded multi-head attention, the model more accurately emphasizes regions with spatial and semantic correlations to the target, enhancing both focus and discrimination for sparse targets. Extensive experiments conducted on two benchmark datasets (SSDD and HRSID) demonstrate the effectiveness and superiority of the proposed method. Overall, the method significantly improves the detection of sparse, multi-scale, and randomly distributed ship targets in SAR images.

Deep learning-based frameworks have shown great potential in the field of hyperspectral image (HSI) classification owing to their superior modeling capabilities. However, the existence of mixed pixels and spectral heterogeneity limits the discriminant performance of the classifier, which makes it impossible to distinguish the mixed spectra effectively in actual scenarios. To address this gap, we propose a subpixel spectral variability network (S(2)VNet) for hyperspectral image classification, which incorporates complete subpixel information and class features modeled by spectral variability and nonlinear mixture characteristics to enhance classification performance. S(2)VNet is capable of extracting endmembers and abundances based on the nonlinear autoencoder (AE) framework and estimating variability parameters by simultaneously considering scaling factors and perturbation terms to ensure accurate endmember construction. The enhanced subpixel fusion module is further designed to automatically integrate three aspects of abundances, spectral cosine correlation information and pixel-level class features to provide a robust joint representation for the classifier. Extensive experiments on four public HSI datasets demonstrate the superiority and generalization of the proposed method when benchmarked with the state-of-the-art methods.

Existing traditional expansion state observers exhibit good tracking performance for constant and low-frequency disturbances. However, their ability to track non-constant disturbances such as ramp and high-frequency harmonics is inadequate. This paper proposes an extended state observer design method based on the internal model principle. This method achieves precise tracking of non-constant disturbances in the system, effectively addressing the issue of disturbance estimation errors in conventional expansion state observers. When applied to control systems, this approach significantly mitigates or suppresses system vibrations caused by non-constant disturbances, thereby enhancing control accuracy. Furthermore, it demonstrates the stability of the controlled system and the active disturbance rejection controller parameters over a wide range of variations. Simulation results indicate that the ADRC controller based on the proposed observer in this paper offers notable advantages, including high tracking accuracy, strong disturbance rejection capability, and good stability, leading to commendable control performance.

Cross-scene image classification aims to transfer prior knowledge of ground materials to annotate regions with different distributions and reduce hand-crafted cost in the field of remote sensing. However, existing approaches focus on single-source domain generalization to unseen target domains, and are easily confused by large real-world domain shifts due to the limited training information and insufficient diversity modeling capacity. To address this gap, we propose a novel multi-source collaborative domain generalization framework (MS-CDG) based on homogeneity and heterogeneity characteristics of multi-source remote sensing data, which considers data-aware adversarial augmentation and model-aware multi-level diversification simultaneously to enhance cross-scene generalization performance. The data-aware adversarial augmentation adopts an adversary neural network with semantic guide to generate MS samples by adaptively learning realistic channel and distribution changes across domains. In views of cross-domain and intra-domain modeling, the model-aware diversification transforms the shared spatial-channel features of MS data into the class-wise prototype and kernel mixture module, to address domain discrepancies and cluster different classes effectively. Finally, the joint classification of original and augmented MS samples is employed by introducing a distribution consistency alignment to increase model diversity and ensure better domain-invariant representation learning. Extensive experiments on three public MS remote sensing datasets demonstrate the superior performance of the proposed method when benchmarked with the state-of-the-art methods.

Deep learning (DL) has been widely applied into hyperspectral image (HSI) classification owing to its promising feature learning and representation capabilities. However, limited by the spatial resolution of sensors, existing DL-based classification approaches mainly focus on pixel-level spectral and spatial information extraction through complex network architecture design, while ignoring the existence of mixed pixels in actual scenarios. To tackle this difficulty, we propose a novel dual-branch subpixel-guided network for HSI classification, called DSNet, which automatically integrates subpixel information and convolutional class features by introducing a deep autoencoder unmixing architecture to enhance classification performance. DSNet is capable of fully considering physically nonlinear properties within subpixels and adaptively generating diagnostic abundances in an unsupervised manner to achieve more reliable decision boundaries for class label distributions. The subpixel fusion module is designed to ensure high-quality information fusion across pixel and subpixel features, further promoting stable joint classification. Experimental results on three benchmark datasets demonstrate the effectiveness and superiority of DSNet compared with state-of-the-art DL-based HSI classification approaches. The codes will be available at https://github.com/hanzhu97702/DSNet, contributing to the remote sensing community.

The performance of a deep learning-based synthetic aperture radar (SAR) automatic target recognition (ATR) model largely relies on the scale and quality of training samples. However, it is time-consuming and expensive to collect sufficient data in practice. Although generative adversarial network (GAN) provides a way for SAR target image generation, existing GAN-based methods cannot confirm what features the generator learns, thus they struggle in generating precise SAR target images. In this article, we propose an angle transformation GAN (ATGAN) that can generate azimuth-controllable SAR target images while preserving the target details. The key idea of our ATGAN is to reframe the generation task from the perspective of image-to-image translation. To this end, ATGAN consists of two modules, a coarse-to-fine generator that aims to learn the angle transformation in the deep feature space, and then, apply it to manipulate the representation of an input SAR target image to generate a new one, while a spectral-normalized patch discriminator that tries to estimate the probability that an input SAR target image is real rather than fake using a patch-averaged strategy. Combining with spatial transformer and adversarial training paradigm, ATGAN can generate precise SAR target images for ATR. Extensive experiments verify the effectiveness of the proposed ATGAN, and our method outperforms the state-of-the-art method qualitatively and quantitatively.

At present, the convolutional neural network (CNN) has been successfully applied in the field of synthetic aperture radar automatic target recognition (SAR-ATR) due to its strong learning ability and automatic hierarchical feature representation. However, the CNN-based methods are good at extracting the local structural features of the target, but are not sufficient in representing the long-range context information between the target and the scene, which restricts the further improvement of the performance of the current SAR-ATR system. To overcome these limitations, in this letter, we propose a novel model named multilevel attention networks (MANets) for SAR target recognition. MANets consider both the local structural features and long-range contextual information simultaneously to improve the representation ability of SAR targets. First, a CNN backbone with five convolutional layers is built to extract multilevel and multiscale convolutional features from the SAR target. Second, these CNN features are fed into a multilevel attention enhancement module (MAEM) to capture the long-range contextual information from spatial, channel, and cross-level attention (CLA) perspectives. Third, a multiscale attention fusion module (MAFM) is designed to fuse and aggregate the multilevel and multiscale features, further enhancing the representation capability of SAR images. Extensively experimental results on the moving and stationary target acquisition and recognition (MSTAR) dataset demonstrate that the proposed MANets achieve state-of-the-art SAR target recognition performance both in the standard operating condition (99.75% at full data volume) and the extended operating condition (97.92% at full data volume) setups. The source code will be released at github.com/Crush0416/MANets.

Abstract The type parameter helps in scattering mechanism analysis and scattering centre identification. However, current approximate solution methods based on spectral estimation are noise‐sensitive and exhibit low accuracy. In this letter, a high‐precision approach for the type parameter based on the alternating direction method of multipliers (ADMM) is proposed. Logarithmic transformation is used to separate the type parameter and the amplitude from the coupling term, and initially obtain the closed‐form solution of the type parameter. Then, the regularization term of l2‐norm is used for denoising. Finally, a joint optimization model based on ADMM is constructed to effectively estimate the type parameter. Simulation results confirm the high accuracy of the proposed approach.

When synthetic aperture radar (SAR) is conducting remote sensing or terrain mapping, its radar beam is inevitably occluded by the variations in the under-test topography. Although back-projection algorithm (BPA) can theoretically directly solve the imaging problems of topography variations that most current SAR imaging algorithms cannot handle, these BPAs only solve the phase focusing of SAR echo signal, and do not consider the mismatch of SAR imaging results caused by topography occlusion. To solve the mis-imaging issue of the occluded area generated by BPA under the case of topography variation, a topography-based BPA (Topo-BPA) is proposed in this paper. Firstly, a new beam occlusion judgment algorithm based on spherical wave assumption is proposed, and its core is depression angle interpolation and depression angle updating. Then, the proposed Topo-BPA embeds the proposed beam occlusion judgment algorithm before the classical BPA, which not only did not reduce the focus depth of BPA, but improved the imaging accuracy of classical BPA. Finally, numerical experiments have demonstrated the superiority of the Topo-BPA’s performance in comparison with classical BPA.

In recent years, deep learning (DL) algorithms have been successfully applied in synthetic aperture radar automatic target recognition (SAR-ATR) owing to its powerful and excellent target feature extraction and representation ability. However, these DL-based models merely exploit the intensity (magnitude) information of SAR target, without fully considering the domain characteristics underlying the SAR images, for example, azimuth, scattering center, phase and so on. To address this issue, this paper proposes a novel information dissemination networks, called IDNets, by both considering the azimuth and strong scatter centers of SAR target in a multi-scale information dissemination mechanism to improve the representation capability of SAR recognition model. Moreover, IDNets introduces a stream-based self-attention (SSA) mechanism to adaptively learn the attention distribution of the multi-streams multi-scale sematic features, further enhancing the performance of SAR-ATR system. Experimental results conducted on the MSTAR dataset demonstrate the effectiveness and superiority of the proposed IDNets compared to the current state-of-the-art DL-based SAR-ATR methods.

Synthetic aperture radar (SAR) images possess diverse domain characteristics that describe targets from multiple perspectives, including scattering center, intensity, phase, azimuth angle, and shadow. The imaging mechanism of SAR images significantly differs from optical images due to its use of radio waves instead of visible light for imaging purposes. However, current deep learning (DL)-based automatic target recognition (ATR) approaches for SAR primarily focus on the intensity information while neglecting the comprehensive consideration of SAR domain characteristics. This limitation results in insufficient expression of target features and hampers further improvement in SAR target recognition performance. To address these issues, this paper proposes a novel SAR target recognition method called KDNets based on information dissemination networks (IDNets) and knowledge hierarchy division (KnHD). KDNets effectively enhances the feature representation ability of SAR targets by exploiting multiple domain characteristics at sample, feature, and decision levels. Specifically, IDNets incorporates scattering center and azimuth angle at sample and feature levels to extract rich multi-scale semantic features for representing SAR targets. At the decision stage, KnHD utilizes azimuth information to rapidly search and fuse prior knowledge about targets from a knowledge base to achieve interpretable and high-precision SAR target recognition. Experimental results on the MSTAR dataset and MA-SAR dataset demonstrate that our proposed KDNets can achieve state-of-the-art performance in SAR target recognition tasks, with the accuracy of 99.75 % in the MSTAR dataset and 98.61 % in the MA-SAR dataset, validating its effectiveness and superiority. Furthermore, this study highlights the significant potential and application prospects of leveraging domain characteristics in SAR-ATR tasks.

In synthetic aperture radar automatic target recognition (SAR-ATR), the limitations of imaging environment and observation conditions make it challenging to acquire a substantial amount of high-value targets, resulting in a severe shortage of datasets. This scarcity leads to poor performance and instability in few-shot SAR target recognition. To address these shortcomings, this article proposes meta-adaptive stochastic gradient descent (Mada-SGD), a novel inner loop parameter update approach based on meta-adaptive hyperparameter learning. By considering the correlation information between multiple update steps, Mada-SGD learns the weight distribution information of initialization parameters across previous and current update steps, akin to a memory mechanism. This approach enhances feature extraction and representation ability for few-shot SAR targets. In addition, an adaptive hyperparameter update strategy is introduced to simultaneously learn the initialization, weight factor, update factor, and update direction in the meta-learner. This effectively resolves parameter updating issues in meta-learning models while improving fast adaptation for few-shot SAR targets. Experimental results on the specialized moving and stationary target acquisition and recognition few-shot learning (MSTAR-FSL) dataset demonstrate that Mada-SGD outperforms the latest few-shot SAR target recognition model in terms of SAR target recognition performance, validating its advancement and superiority.

At present, due to the limitations of the imaging environment and observation conditions, the automatic target recognition of synthetic aperture radar (SAR-ATR) encounters a severe shortage of target samples, which leads to poor recognition and unstable performance for few-shot targets. To address the above issues, this paper proposes an inner-loop parameter update method based on meta-adaptive hyper-parameter learning, called Mada-SGD, to achieve the goal of efficient recognition of few-shot SAR targets. In Mada-SGD, an adaptive hyper-parameter update strategy is introduced to automatically learn the initialization, weight factor, update factor and update direction in the meta-learner, it effectively solves the problem of parameter update in the meta-learning model and improves the fast adaptation of few-sample SAR targets. In addition, Mada-SGD learns the weight distribution information of initialization parameters by fully considering the correlation information between multi-step updates, which is similar to a memory mechanism and improves the feature extraction and representation ability of few-shot SAR targets. The experimental results on the customized MSTAR dataset show that the proposed Mada-SGD is able to achieve the state-of-the-art few-shot SAR target recognition performance, which verifies its effectiveness and reliability.

Accurate crop mapping is of great significance for crop yield forecasting, agricultural productivity development and agricultural management. Thanks to its all-time and all-weather capability, integrating multi-temporal synthetic aperture radar (SAR) for crop mapping has become essential and challenging task in remote sensing. In recent years, deep learning (DL) has demonstrated excellent crop mapping accuracy to interpret crop dynamics. However, existing DL-based methods tend to be incapable of capturing spatial and temporal features at different scales simultaneously, and this often leads to severe mis-classification due to the complex and heterogeneous distribution of crops and diverse phenological patterns. In this paper, we propose a novel spatio-temporal multi-level attention method, named as STMA, for crop mapping using time-series SAR imagery in an end-to-end fashion to increase the capability of crop phenology retrieval. Specifically, the multi-level attention mechanism is designed to aggregate multi-scale spatio-temporal representations on crops via cascaded spatio-temporal self-attention (STSA) and multi-scale cross-attention (MCA) modalities. To ensure a fine extraction of multi-granularity features, a learnable spatial attention position encoding is proposed to adaptively generate the position priors to facilitate multi-level attention learning. Experimental results on Brandenburg Sentinel-1 dataset, public PASTIS-R dataset and South Africa dataset demonstrated that STMA can achieve state-of-the-art performance in crop mapping tasks, with the accuracy of 96.54% in the Brandenburg Sentinel-1 dataset, 86.77% in the PASTIS-R dataset and 83.37% in the South Africa dataset, validating its effectiveness and superiority. Further comparison of spatio-temporal generalization capability reflected its excellent performance in spatio-temporal modeling on different crops and scenarios. This research provides a viable and intelligent spatio-temporal framework for large-area crop mapping using time-series SAR imagery in complex agricultural systems. The Brandenburg Sentinel-1 dataset and the STMA code will be publicly available at https://github.com/hanzhu97702/ISPRS_STMA.

Synthetic aperture radar (SAR) image similarity metric is at the core of SAR image interpretation techniques, however, it is still a challenging task due to complex nonlinear intensity, scale, and rotation differences between SAR images and other remote sensing images. This letter addresses this problem by proposing a novel similarity metric method for SAR images using structure and shape properties. The magnitude and orientation representation of the phase congruency model is first built based on the local phase of images. Then a new scale and rotation-invariant local binary pattern (SRI-LBP) descriptor is proposed using local structure and shape information. Finally, a similarity metric is defined using the symmetry Kullback Leibler divergence (SKLD) of the SRI-LBP descriptors. Numerical experiment results verify its robustness in terms of nonlinear intensity, scale, and rotation differences.

Synthetic aperture radar raw data simulation (SAR-RDS) is beneficial to the SAR system design, signal processing method verification, and radar parameter optimization. Most SAR-RDS methods are based on the flat terrain assumption. However, the fluctuant terrain in real scene will induce severe SAR beam occlusion effect and produce radar shadow, leading to incorrect RDS results. Thus, a dynamic elevation angle interpolation (DEAI) algorithm is proposed for SAR shadow judgment by considering the actual SAR working process. The key of the proposed DEAI algorithm is the 1-D EAI and shadow visualization update, which avoids the problem that the existing methods cannot judge the shadow of partial areas due to the insufficiently refined mesh grid or the mismatch of the judgment model. Moreover, an evaluation framework named as joint image and signal criteria (JISC) is proposed from the perspectives of SAR imaging and signal processing results to objectively evaluate the SAR-RDS results and solve the problem that the existing evaluation methods cannot be compatible with fluctuant terrain. Finally, the numerical experiment verified our theoretical analyses.

In recent years, deep learning (DL) has achieved excellent results in synthetic aperture radar automatic target recognition (SAR-ATR) tasks. However, most of the existing SAR target recognition methods with DL are based on real-valued calculations (only using magnitude information), which ignore the phase information of the target, resulting in poor target recognition. In order to take full advantage of the phase information, this paper proposes an adaptive multi-stream feature fusion SAR target recognition method based on complex-valued calculations, called AdaMS-CVNet, to build an end-to-end complex-valued SAR target recognition framework. In the beginning, a multi-stream structure based on complex-valued calculation is introduced to obtain a wealth of multi-scale feature information of the SAR target. Besides, to enhance the representation ability of targets, an attention mechanism is employed to effectively fuse the multi-stream complex-valued features. Finally, the experimental results on the Complex-MSTAR dataset illustrate that the proposed AdaMS-CVNet can achieve better SAR target recognition performance than the most advanced real-valued calculation models, which verifies its effectiveness.

Bearing is a key part of rotating machinery. Accurate prediction of bearing life can avoid serious failures. To address the current problem of low accuracy and poor predictability of bearing life prediction, a bearing life prediction method based on digital twins is proposed. Firstly, the vibration signals of rolling bearings are collected, and the time-domain and frequency-domain features of the actual data set are extracted to construct the feature matrix. Then unsupervised classification and feature selection are carried out by improving the self-organizing feature mapping method. Using sensitive features to construct a twin dataset framework and using the integrated learning CatBoost method to supplement the missing data sets, a complete digital twin dataset is formed. Secondly, important information is extracted through macro and micro attention mechanisms to achieve weight amplification. The life prediction of rolling bearing is realized by using fusion features. Finally, the proposed method is verified by experiments. The experimental results show that this method can predict the bearing life with a limited amount of measured data, which is superior to other prediction methods and can provide a new idea for the health prediction and management of mechanical components.