Magnetic anomaly detection (MAD) is an all-weather passive ferromagnetic target detection technique with high environmental stability. However, in the absence of prior target information, the detection performance of existing MAD methods remains to be improved under low signal-to-noise ratio (SNR) conditions. To address the bottleneck, this study proposes a high-performance intelligent MAD method based on stochastic resonance (SR) theory. First, this study modifies a more flexible potential function to better fit the measured magnetic anomaly signals. Second, the enhancement mechanism of second-order underdamped SR on weak magnetic anomaly signals is systematically investigated. Finally, an intelligent parameter optimization system is established by integrating the quantum-behaved particle swarm optimization (QPSO) algorithm to solve the long-standing parameter tuning difficulty of SR systems. Experimental verification shows that, under a 2% false alarm rate, the proposed method achieves a detection rate of 86.07% in a -5 dB environment, which is 23.4% higher than that of the parallel stochastic resonance (PSR) method. This method solves the problem of difficult parameter adjustment in SR systems, achieves a breakthrough in detection rate, and significantly enhances the application prospects of SR methods in practical scenarios.

Segmenting key components of space targets using Inverse Synthetic Aperture Radar (ISAR) images is an important interpretation task in space situational awareness. However, the scarcity of pixel-level annotated data, inter-class confusion caused by morphological differences among multiple target classes, and the absence of structural priors for components restrict the performance improvement in existing deep models on this task. Therefore, this paper proposes a Semi-Supervised Structural Prior-Guided Network (SSPNet). First, a Gated Manifold-Constrained Hyper-Connections Vision Transformer (GMHC-ViT) encoder is proposed to broaden the feature representation space via parallel multi-feature streams with adaptive gating, thereby alleviating inter-class confusion and enhancing cross-category generalization. Second, a Prior-Guided Module (PGM) is proposed to extract shape and edge priors of components, and it adaptively enhances the weakly activated channels of encoder features through cross-attention, thereby injecting structural knowledge independent of image quality into the segmentation process. Furthermore, to effectively leverage large amounts of unlabeled data, a strong perturbation strategy tailored to the characteristics of ISAR images is designed for consistency regularization. Experimental results on a simulated ISAR dataset containing 38 classes of space targets demonstrate that SSPNet outperforms existing methods and exhibits strong segmentation capability even under low signal-to-noise ratio (SNR) conditions.

Three-dimensional (3-D) InISAR imaging plays an essential role in ensuring spacecraft reliability. However, spacecraft 3-D rotation introduces space-time varying Doppler during imaging, complicating simultaneous image focusing and interferometric phase error mitigation. This article proposes a multichannel joint azimuth compression InISAR imaging algorithm based on local region-of-interest (ROI) processing. To tackle decorrelation phase errors, an overall InISAR imaging framework based on multichannel joint compensation is designed. In this framework, each motion compensation step is performed using consistent parameters across all channels, ensuring strict coherence. Furthermore, to prevent accumulated phase errors, an ROI-based multichannel joint processing method is proposed. Otsu's thresholding and rectangular coverage optimization are employed for ROI extraction and segmentation, with the resulting subimages serving as independent signal processing units. For focusing the ROI subimages, a local fractional Fourier transform (FrFT)-based azimuth compression method is introduced. The technique leverages the Local FrFT to differentially compensate quadratic phase terms and limits computational complexity by exponential decay search. Simulation and electromagnetic computational data demonstrate that the proposed algorithm effectively focuses ISAR images and reduces interferometric phase errors, yielding superior 3-D imaging accuracy compared to conventional InISAR imaging methods.

In ground-based space surveillance, space target attitude estimation is critical for space situational awareness, yet existing methods based on inverse synthetic aperture radar (ISAR) images suffer from three core limitations: phase information is discarded in amplitude-only processing, convolutional neural networks have a restricted global receptive field, and the physical topology of satellite components is not explicitly modeled. To address these issues, we propose a complex-domain Transformer–graph neural network (CD-TrGNN) that unifies global context modeling and adaptive topological reasoning in an end-to-end framework. Specifically, a complex-domain Transformer module (CD-Transformer) with tailored attention captures long-range dependencies among image patches while preserving both amplitude and phase information; a complex-domain graph convolution module (CD-GC) with learnable adjacency matrices and a dual-path update mechanism explicitly encodes the structural relationships among satellite parts. On a self-built ISAR complex image dataset, CD-TrGNN achieves a three-axis mean absolute error of only 1.70°, substantially outperforming six representative baselines. Ablation experiments confirm the effectiveness of complex-domain processing, global attention, and topological reasoning. At a 5 dB signal-to-noise ratio, the error remains at 2.81°, and the accuracy stays below 2° for two different satellite structures. These results demonstrate that CD-TrGNN can fully exploit the information in ISAR complex images, enabling high-accuracy and highly robust attitude estimation.

Semantic segmentation technology based on Inverse Synthetic Aperture Radar (ISAR) images can provide crucial perception and analytical capabilities for intelligent safety maintenance of on-orbit spacecraft. However, conventional semantic segmentation methods suffer from three main limitations: firstly, the lack of modeling for radar physical characteristics in the “image first, segment later” pipeline leads to loss of scattering information and phase details; secondly, reliance on extensive pixel-level manual annotation increases application costs; thirdly, ineffective utilization of spacecraft structural priors fails to guide networks to focus on the main body and edges of spacecraft segmentation. To address these issues, this paper proposes a complex-domain semantic segmentation framework named One-Stop Segmentation (OSS) based on ISAR echoes. The framework incorporates two innovative modules: an Automatic ISAR Labeling (AIL) method designed based on ISAR scattering characteristics to generate labels corresponding to ISAR echoes, and a complex-domain semantic segmentation network named One-Stop Segmentation Network (OSSNet) that performs semantic segmentation directly on echoes, avoiding information loss from imaging while shortening the data processing chain. Core contributions of OSSNet include: (1) a Domain Alignment Module (DAM) to effectively mitigate domain mismatch caused by data distribution differences between raw echo signals and labels; (2) a Multi-Perspective Attention (MPA) framework incorporating a Sliding Correlation Attention (SCA) module and a Subdomain Balanced Attention (SBA) module, lever-aging spacecraft structural priors to guide the network’s focus on main structures and edge details from complementary perspectives, significantly improving segmentation ac-curacy. Experimental results on a simulated ground-based radar dataset demonstrate that the proposed OSS framework achieves a mean Intersection over Union (mIoU) of 92.13% and a mean F1-score of 95.75% in ISAR spacecraft semantic segmentation tasks, outperforming existing methods.

Complex natural resonance frequency is only determined by target's intrinsic properties, and independent of incident angle and target's attitude, making target recognition based on complex natural resonance frequency a promising solution. However, the existing extraction methods inevitably introduce calculation errors, resulting in inaccurate extraction of it, which limits the improvement of recognition accuracy. To address these issues, physics-informed complex natural resonance frequency correction is proposed. First, a loss function based on target's resonance scattering mechanism is designed. Physics is used to drive network to extract features, strengthening the physical constraints on the corrected data. Second, by introducing the idea of complex-valued neural network, a complex domain-variational autoencoder is designed, which considers the correlation between the real and imaginary parts of complex data, thereby excavating more internal features of data. Experimental results verify the effectiveness of the proposed method. Compared with the original data, the relative errors are reduced by 18.02% for the real part and 0.49% for the imaginary part. Improving the consistency between corrected data and target's resonance scattering mechanism.

Aiming at the problems of low recognition accuracy in Inverse Synthetic Aperture Radar (ISAR) satellite image recognition, such as imaging distortion and defocus caused by target attitude changes and non-uniform rotation, scarce samples, and insufficient feature extraction capabilities, this paper proposes a residual network model YN-Mish-CBAM-SD that integrates multi-dimensional improvement strategies for ISAR image satellite target recognition. This paper construct a dataset containing distorted and defocused ISAR image satellite targets. Based on the ResNet network, the Mish activation function is introduced to replace the traditional ReLU activation function to enhance the network's nonlinear expression capabilities. The Convolutional Block Attention Module (CBAM) is combined to dynamically allocate attention weights in the channel and spatial dimensions, significantly improving the perception of target structures in ISAR images. Meanwhile, Stochastic Depth is embedded in the residual blocks to alleviate overfitting by probabilistically skipping partial network layers, enhancing the model's generalization ability. The proposed model effectively extracts deep discriminative features of ISAR images while maintaining computational efficiency. Testing experiments on the constructed satellite image dataset show that the improved network achieves an average recognition accuracy of 72.9% in ISAR image satellite target recognition and classification tasks, which is an increase of 55.2%, 39.9%, and 46.4% compared to the basic ResNet18, ResNet50, and ResNet101 networks, respectively.

Enhancing generalization capabilities and robustness in scenarios with limited sample sizes, while simultaneously decreasing reliance on extensive and high-quality datasets, represents a significant area of inquiry within the domain of radar target recognition. This study introduces a few-shot learning framework that leverages multimodal feature fusion. We develop a cross-modal representation optimization mechanism tailored for the target recognition task by incorporating natural resonance frequency features that elucidate the target’s scattering characteristics. Furthermore, we establish a multimodal fusion classification network that integrates bi-directional long short-term memory and residual neural network architectures, facilitating deep bimodal fusion through an encoding-decoding framework augmented by an energy embedding strategy. To optimize the model, we propose a cross-modal equilibrium loss function that amalgamates similarity metrics from diverse features with cross-entropy loss, thereby guiding the optimization process towards enhancing metric spatial discrimination and balancing classification performance. Empirical results derived from simulated datasets indicate that the proposed methodology achieves a recognition accuracy of 95.36% in the 5-way 1-shot task, surpassing traditional unimodal image and concatenation fusion feature approaches by 2.26% and 8.73%, respectively. Additionally, the inter-class feature separation is improved by 18.37%, thereby substantiating the efficacy of the proposed method.

As a sub field of machine learning, deep learning has shown outstanding benefits in several areas compared to more traditional, non-deep machine learning. However, continuous expansion of learning cannot be achieved by relying solely on existing deep learning methods. Therefore, this paper proposes a deep learning-based IFL (instant feature learning) algorithm for radar target recognition. The Adaptive Continual Memory (ACM) algorithm is used as the main method, supplemented by HOG feature extraction algorithm and HNSW index optimization algorithm, to carry out initial segmentation and learning of the dataset, and to carry out continuous learning, training, and testing based on the data stream, to achieve the effect of instant feature learning. The experiments of the IFL algorithm for radar target recognition based on deep learning in the field of radar target recognition verified its breakthrough progress in instant feature learning, especially in the satellite target recognition task achieved 99.86% accuracy, fully proving the practical value of the algorithm.

The application of large-angle inverse synthetic aperture radar (ISAR) imaging significantly enhances resolution and facilitates the capture of more intricate target details. However, 3-D larger angles can amplify previously negligible motion errors, complicating the imaging process. This article introduces a nonuniform polar format and fractional Fourier transform segmentation compensation (NUPF-FRFTSC) ISAR imaging algorithm, based on a high-precision large-angle signal model. The algorithm efficiently decouples motion errors into distinct 2-D and 3-D components, sequentially correcting each. To address the complex high-order 2-D spatially varying (SV) range cell migration (RCM) and phase error (PE), the NUPF algorithm (NUPFA) is proposed. This approach mitigates 2-D errors by mapping echo data acquired during nonuniform rotation into a uniform wavenumber domain. The necessary rotation parameters for this mapping are estimated through a combination with image entropy evaluation and particle swarm optimization (PSO). Furthermore, to correct 3-D SV PE, which are independent of the range-Doppler (RD) scale in ISAR images, the FrFTSC method is developed. This technique compensates signals within each wavenumber domain cell regionally, utilizing the similarity of second-order phase coefficients of neighboring scattering points. Electromagnetic computational data and measured experiment demonstrate that the proposed NUPF-FRFTSC algorithm successfully achieves large-angle focused imaging of targets, yielding results superior to traditional ISAR imaging methods.

Inverse Synthetic Aperture Radar (ISAR) images of complex targets have a low Signal-to-Noise Ratio (SNR) and contain fuzzy edges and large differences in scattering intensity, which limits the recognition performance of ISAR systems. Also, data scarcity poses a greater challenge to the accurate recognition of components. To address the issues of component recognition in complex ISAR targets, this paper adopts semantic segmentation and proposes a few-shot semantic segmentation framework fusing multimodal features. The scarcity of available data is mitigated by using a two-branch scattering feature encoding structure. Then, the high-resolution features are obtained by fusing the ISAR image texture features and scattering quantization information of complex-valued echoes, thereby achieving significantly higher structural adaptability. Meanwhile, the scattering trait enhancement module and the statistical quantification module are designed. The edge texture is enhanced based on the scatter quantization property, which alleviates the segmentation challenge of edge blurring under low SNR conditions. The coupling of query/support samples is enhanced through four-dimensional convolution. Additionally, to overcome fusion challenges caused by information differences, multimodal feature fusion is guided by equilibrium comprehension loss. In this way, the performance potential of the fusion framework is fully unleashed, and the decision risk is effectively reduced. Experiments demonstrate the great advantages of the proposed framework in multimodal feature fusion, and it still exhibits great component segmentation capability under low SNR/edge blurring conditions.

Magnetic anomaly detection (MAD) technology, which identifies concealed ferromagnetic targets by analyzing weak perturbations in the geomagnetic field, holds significant value in unexploded ordnance (UXO) identification, underwater target detection, and related fields. To address the performance limitations of existing nonprior detection methods in low signal-to-noise ratio (SNR) scenarios, this study establishes a novel enhanced MAD framework through comparative analysis of probability density functions (pdfs) between reference signals and target signals, coupled with multimetric similarity assessment, enabling reliable MAD in both single-sensor and dual-sensor configurations. Theoretical analysis and field experiments show that at $\text {SNR} = -4$ dB, the proposed method can improve the detection probability by 12.7% and 26.2%, respectively, relative to the parallel stochastic resonance (PSR) method and the minimum entropy detection (MED) method. This approach transcends the information dimensionality constraints of traditional entropy features, establishing a new paradigm for prior-free MAD in complex magnetic environments through multiscale probability density feature extraction.

Here, a target recognition method of natural resonant complex frequency under low Signal-to-Noise Ratio (SNR) is proposed, mainly to solve the problem that natural resonant frequency is greatly affected by noise while the recognition accuracy is low. This method designs a recognition approach for natural resonant complex frequency by studying the characteristics of natural resonant complex frequency and integrating high-order spectrum analysis. Through theoretical analysis of the feasibility of natural resonant complex frequency bispectral transformation, the introduction of the natural resonance decay factor into target recognition is pioneered, providing a novel processing approach for identifying natural resonant complex frequencies.

While the wavenumber-domain approach enables large-angle inverse synthetic aperture radar (ISAR) cross-range scaling, its practical application remains constrained by the target’s non-uniform rotation and scene center offset (SCO). In response to this issue, this paper introduces a novel large-angle ISAR cross-range scaling method through a joint estimation method based on the wavenumber domain. A non-uniform rotational wavenumber-domain signal model with SCO is developed. Utilizing this model and the sensitivity of wavenumber-domain imaging to SCO, a joint estimation algorithm that combines particle swarm optimization (PSO) and image entropy evaluation is proposed, achieving accurate parameter estimation. Leveraging the estimated parameters, the range and cross-range scaling factors in the wavenumber-domain imaging are derived, facilitating ISAR cross-range scaling with higher accuracy than the traditional method. The effectiveness and robustness of the proposed method are validated under various conditions, through scattering point and electromagnetic computing simulation.

The application of deep learning to inverse synthetic aperture radar (ISAR) target recognition helps to improve accuracy in space target monitoring. However, the orbit transfer and maneuver of space targets are likely to cause range migration in radar echoes. Information deficiency of scattering points and the limitations of data acquisition methods pose enormous challenges to space target imaging and recognition. To address these issues, this paper proposes a semi-supervised space target recognition algorithm based on an integrated network of imaging and recognition in the radar signal domain. By directly processing radar signals, the algorithm can achieve high-precision space target imaging and recognition under the general situation and the conditions of range migrations. Based on the inherent characteristics of radar complex echoes, the algorithm utilizes unlabeled echoes to generate pseudo labels to achieve better generalization capabilities. Both real and complex convolutions are exploited to generate high-resolution features. Besides, feature optimization modules are designed to effectively integrate high-resolution texture features and contour features to magnify the difference between the target and background. The ablation experiments and contrast experiments indicate that, under the conditions of migration and missing components, the algorithm can obtain high-resolution features for imaging and recognition by only using a small amount of labeled data. Thus, the algorithm achieves high accuracy and robustness for migrated space target recognition and has superiority over other algorithms.

The continuous emergence of new targets in open scenarios leads to a substantial decrease in the performance of Inverse Synthetic Aperture Radar (ISAR) recognition systems. Also, data scarcity further exacerbates the challenge of identifying new classes of ISAR targets. In this paper, a few-shot incremental target recognition framework based on Scattering-Topology Properties (STPIL) is proposed. Specifically, STPIL extracts scattering-topology properties of ISAR targets as recognition features. Meanwhile, the pseudo-incremental training strategy effectively alleviates the algorithm’s forgetting of old knowledge, and improves compatibility with new classes. Besides, a feature embedding network, with few parameters, is designed based on the graph neural network. This embedding network is highly adaptable to changes in data distribution. Additionally, STPIL fully considers the joint distribution and marginal distribution in scattering features, and uses the Brownian distance metric module to make the scattering-topology features more discriminative. Experimental results on both the simulation dataset and the public measured data indicate that STPIL can effectively balance new classes with old classes, and has superior performance to other advanced methods in the incremental recognition of targets.

As target recognition has become a research hot-spot in the space field, radar imaging as the front-end technology of target recognition has become the key to identification, and the simulation technology of radar echo is the focus of research. Aiming at the difficulty of obtaining open source data of radar echo and the complicated problem of establishing dynamic target scene with MATLAB, this paper proposes a new method of dynamic target radar echo simulation by using the joint simulation technology of MATLAB control CST. This method can provide the echo research data closer to the real situation for the theoretical study of the spatial object recognition, which is of great significance for the subsequent depth study of the radar imaging technology.

This paper addresses the issues of short impulse response signal duration and susceptibility to noise interference by using the Leaky Integrate-and-Fire (LIF) neuron model to simulate the charging and discharging process for impulse signal generation. In addition, an Impulse Response Signal-Convolutional Spiking Neural Network (IRS-CSNN) is developed to exploit the time-frequency characteristics of the target impulse response signal for target identification. First, the signal sequence of pitch and azimuth angle changes is subjected to time-frequency analysis, from which the time-frequency features are extracted and fed into the IRS-CSNN with Time-Frequency Spatial Attention Mechanism (TF-SAM) to facilitate target classification and identification. The experimental results show that this approach effectively utilizes the feature information in the impulse response signal and exploits the feature extraction capabilities of the CSNN. The average recognition accuracy for targets at different angles under noise-added conditions is 89.93%, an improvement of 1.87%-4.34% over that of a similarly structured CNN and other machine learning models. These results confirm the effectiveness of the proposed method.

Synthetic Aperture Radar (SAR) is renowned for its all-weather and all-time imaging capabilities, making it invaluable for ship target recognition. Despite the advancements in deep learning models, the efficiency of Convolutional Neural Networks (CNNs) in the frequency domain is often constrained by memory limitations and the stringent real-time requirements of embedded systems. To surmount these obstacles, we introduce the Split_ Composite method, an innovative convolution acceleration technique grounded in Fast Fourier Transform (FFT). This method employs input block decomposition and a composite zero-padding approach to streamline memory bandwidth and computational complexity via optimized frequency-domain convolution and image reconstruction. By capitalizing on FFT’s inherent periodicity to augment frequency resolution, Split_ Composite facilitates weight sharing, curtailing both memory access and computational demands. Our experiments, conducted using the OpenSARShip-4 dataset, confirm that the Split_ Composite method upholds high recognition precision while markedly enhancing inference velocity, especially in the realm of large-scale data processing, thereby exhibiting exceptional scalability and efficiency. When juxtaposed with state-of-the-art convolution optimization technologies such as Winograd and TensorRT, Split_ Composite has demonstrated a significant lead in inference speed without compromising the precision of recognition.