Sparse ISAR Imaging Method Based on Self-Attention Dictionary Learning

HU Changyu; WANG Yu; CHENG Yuan

doi:10.11959/j.issn.1000-436x.TXXB250634

您当前的位置：

首页 >

文章列表页 >

Sparse ISAR Imaging Method Based on Self-Attention Dictionary Learning

更新时间：2026-05-13

- Sparse ISAR Imaging Method Based on Self-Attention Dictionary Learning
- Journal on Communications (2026)
- 作者机构：
  
  1.无锡学院电子信息工程学院，江苏无锡 214000
  2.南京航空航天大学雷达成像与微波光子技术教育部重点实验室，江苏南京 210000
  3.中国航空工业集团雷华电子技术研究所，江苏无锡 214000
- 作者简介：
- 基金信息：
  
  The National Natural Science Foundation(62501274);The General Program of Basic Science (Natural Science) Foundation of the Jiangsu Higher Education Institutions(23KJB510035);The Key Laboratory of Radar Imaging and Microwave Photonics Technology (Nanjing University of Aeronautics and Astronautics), Ministry of Education(NJ20230006);Wuxi University Research Start-up Fund for Introduced Talents(2023r014)
- DOI：10.11959/j.issn.1000-436x.TXXB250634
  CLC： TN95
- Received：26 November 2025，
  
  Revised：2026-05-11，
  
  Accepted：11 May 2026，
- 稿件说明：
移动端阅览
HU Changyu, WANG Yu, CHENG Yuan. Sparse ISAR Imaging Method Based on Self-Attention Dictionary Learning[J/OL]. Journal on Communications, 2026.
DOI：

HU Changyu, WANG Yu, CHENG Yuan. Sparse ISAR Imaging Method Based on Self-Attention Dictionary Learning[J/OL]. Journal on Communications, 2026. DOI： 10.11959/j.issn.1000-436x.TXXB250634.

摘要

稀疏逆合成孔径雷达(Inverse Synthetic Aperture Radar

ISAR)成像方法能在有限观测条件下重建出图像对比度高、虚假散射点少的目标图像。结合稀疏变换字典的稀疏ISAR成像方法能够进一步提升目标轮廓和散射点分布的重建质量。然而，在低压缩比条件下，此类方法仍易出现目标轮廓重建不完整的问题。其根本原因在于稀疏变换字典所提取的稀疏表示仅能刻画目标的局部结构特征与散射分布特性，难以充分表征目标的全局结构特征与散射分布特性。为增强稀疏表示对全局结构特征的表征能力，本文引入自注意力机制(Self-Attention Mechanism

SAM)以指导字典学习(Dictionary Learning

DL)，使得稀疏变换字典能够从测量数据中充分学习目标的全局结构特征。结合SAM的DL模型称为SAM-DL，进一步提出基于SAM-DL的稀疏ISAR成像方法，用于低压缩比条件下目标成像。成像结果表明，相较于现有的DL类成像方法，SAM-DL成像方法在低压缩比条件下重建的目标轮廓更完整，成像性能更优越。

Abstract

Sparse inverse synthetic aperture radar (ISAR) imaging methods are capable of reconstructing target images with high contrast and few artifacts under limited observation conditions. Incorporating sparse transform dictionaries into sparse ISAR imaging can further enhance the reconstruction quality of target contours and scattering distributions. However

these methods still tend to produce incomplete target contours under conditions of low compression ratios. This limitation arises because the sparse representations derived from sparse transform dictionaries primarily capture local structural or scattering features of the target

while failing to sufficiently represent the global structural and scattering features of the target. To strengthen the modeling of global structures within sparse representations

we introduce a self-attention mechanism (SAM) to guide dictionary learning (DL)

enabling the sparse transform dictionary to effectively capture global structural features of target from under-sampled measurements. The resulting DL model is referred to as SAM-DL

and we further develop the SAM-DL based sparse ISAR imaging method for target reconstruction under low compression ratios. Experimental results demonstrate that

compared with existing DL-based imaging approaches

the SAM-DL method is able to reconstruct more complete target contours under low compression ratios

achieving superior imaging performance.

关键词

Keywords

references

张群 , 张宏伟 , 倪嘉成 , 等 . 合成孔径雷达深度学习成像研究综述 [J ] . 信号处理理 , 2023 , 39 ( 09 ): 1521 - 1551 .

ZHANG Q ， ZHANG H W ， NI J C ， et al . A survey of synthetic aperture radar imaging methods based on deep learning [J ] . Journal of Signal Processing , 2023 , 39 ( 9 ): 1521 - 1551 .

WU Y , LI R , MAI Y , et al . Rr-admm-net for sparse aperture ISAR imaging under low signal-to-noise ratio [J ] . IEEE Transactions on Aerospace and Electronic Systems , 2026 . (Early Acess)

LV M , YANG J , WANG D , et al . A structural sparse ISAR imaging method with joint phase autofocusing [J ] . IEEE Geoscience and Remote Sensing Letters , 2024 , 21 : 1 - 5 .

CETIN M , KARL W M . Feature-enhanced synthetic aperture radar image formation based on non-quadratic regularization [J ] . IEEE Transactions on Image Processing , 2001 , 10 ( 4 ): 623 – 631 .

HUANG Q , ZHANG L , WEI S , et al . Learned 2D-TwISTA for 2-D sparse ISAR imaging [J ] . IEEE Geoscience and Remote Sensing Letters , 2025 , 22 : 1 - 5 .

WANG L Z , WANG L , CONDE M H , et al . A fast total-variation driven network for sparse aperture ISAR imaging [J ] . Digital Signal Processing , 2024 , 151 : 104515

全英汇 , 吴耀君 , 邢孟道 , 等 . 基于稀疏恢复的雷达信号处理研究综述 [J ] . 雷达学报 , 2024 , 13 ( 01 ): 46 - 67 .

QUAN Y H , WU Y J , XING M D , et al . A review of radar signal processing based on sparse recovery [J ] . Journal of Radars , 2024 , 13 ( 01 ): 46 – 67 .

YANG J , WANG Y . Sparse aperture inverse synthetic aperture radar imaging for maneuvering targets with migration through resolution cells correction [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2025 , 18 : 4446 - 4472 .

DONOHO D . Compressed sensing [J ] . IEEE Transactions on Information Theory , 2006 , 52 ( 4 ): 1289 - 1306 .

ZHANG Y , BAI X , ZHOU F . High-resolution ISAR imaging of maneuvering targets based on 2-D complex fast sparse Bayesian learning [J ] . IEEE Transactions on Radar Systems , 2025 , 3 : 995 - 1005 .

WANG L , LOFFELD O , MA K L , et al . Sparse ISAR imaging using a greedy kalman filtering approach [J ] . Signal Processing , 2017 , 138 ( C ): 1 - 10 .

胡长雨 , 汪玲 , 朱栋强 . 结合字典学习技术的ISAR稀疏成像方法 [J ] . 电子与信息学报 , 2019 , 41 ( 7 ): 1735 - 1742 .

HU C Y , WANG L , ZHU D Q . Sparse ISAR imaging exploiting dictionary learning [J ] . Journal of Electronics & Information Technology , 2019 , 41 ( 7 ): 1735 - 1742 .

HU C Y , WANG L , ZHU D Y , et al . Inverse synthetic aperture radar sparse imaging exploiting the group dictionary learning [J ] . MDPI Remote Sensing , 2021 , 13 ( 14 ): 2812 - 2833 .

胡长雨 , 陈春风 , 汪玲 , 等 . 结合最近邻图模型的稀疏ISAR成像方法 [J ] . 电子学报 , 2024 , 52 ( 01 ): 170 - 180 .

HU C Y , CHEN C F , YI W Y , et al . Sparse ISAR imaging combined with nearest neighbor graph model [J ] . Acta Electronica Sinica , 2024 , 52 ( 01 ): 170 - 180 .

ZHANG C , ZHANG S , LIU Y , et al . Fast bayesian method for joint sparse ISAR imaging and motion compensation for uniform rotating targets [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2024 , 62 : 1 - 18 .

AHARON M , ELDA M , BRUCKSTEIN A . K-SVD: An algorithm for designing overcomplete dictionaries for sparse representation [J ] . IEEE Transactions on Signal Processing , 2006 , 54 ( 11 ): 4311 - 4322 .

VASWANI A , SHAZEER N , PARMAR N , et al . Attention is all you need [C ] // Proceedings of the 31st Conference on Neural Information Processing Systems (NIPS 2017 ): ACM Press , 2017 : 1 - 11 .

ZHANG Z , WU Y , ZHOU J , et al . SG-Net: syntax guided transformer for language representation [J ] . IEEE Transactions on Pattern Analysis and Machine Intelligence , 2022 , 44 ( 6 ): 3285 - 3299 .

JARQUIN-VASQUEZ H , ESCALANTE H J , MONTES-Y-GOMEZ M , et al . GHA: A gated hierarchical attention mechanism for the detection of abusive language in social media [J ] . IEEE Transactions on Affective Computing , 2025 , 16 ( 2 ): 946 - 959 .

ZHANG X , ZHOU Y , YANG G , et al . Anchor attention, small cache: code generation with large language models [J ] . IEEE Transactions on Software Engineering , 2025 , 51 ( 6 ): 1866 - 1881 .

ZAFAR A . Single stage adaptive multi-attention network for image restoration [J ] . IEEE Transactions on Image Processing , 2024 , 33 : 2924 - 2935 .

JIANG B , LIU Y , CHEN X , et al . Graph attention in attention network for image denoising [J ] . IEEE Transactions on Systems, Man, and Cybernetics: Systems , 2023 , 53 ( 11 ): 7077 - 7088 .

PEI , X , DENG X , XIONG N N , et al . Complex graph analysis and representation learning: problems, techniques, and applications [J ] . IEEE Transactions on Network Science and Engineering , 2024 , 11 ( 5 ): 4990 - 5007 .

ZHU D Y , WANG L , YU Y S , et al . Robust ISAR range alignment via minimizing the entropy of the average range profile [J ] . IEEE Geosciences and Remote Sensing Letters , 2009 , 6 ( 2 ): 204 – 208 .

汪玲 , 朱岱寅 , 朱兆达 . 基于SAR实测数据的舰船成像研究 [J ] . 电子与信息学报 , 2007 , 29 ( 2 ): 401 - 404 .

WANG L , ZHU D Y , ZHU Z D . Study on ship imaging using SAR real data [J ] . Journal of Electronics and Information Technology , 2007 , 29 ( 2 ): 401 - 404 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Grid self-attention mechanism 3D object detection method based on raw point cloud

Nonparametric Bayesian dictionary learning algorithm based on structural similarity

Incoherent dictionary learning and sparse representation for single-image rain removal

Speech enhancement based on multi-task sparse representation for dual small microphone arrays

Gesture recognition approach based on learning sparse representation

Related Author

Yang SUN

Zhenyu YANG

Bin LU

Ge LIU

Jian KANG

Wenbiao TIAN

Guosheng RUI

Daoguang DONG

Related Institution

School of Control and Compute Engineering, North China Electric Power University

Engineering Research Center of Intelligent Computing for Complex Energy Systems, Ministry of Education

Signal and Information Processing Key Laboratory in Shandong,Navy Aviation University

College of Information Engineering, Xiangtan University

College of Electrical and Information Engineering,Hunan University

AI问答

⁰