基于多分支多信息多深度复值特征融合网络的SAR舰船目标识别方法
Recognition Method of Ship Targets for SAR Based on M3Net
- 电子学报 2025年53卷第10期 页码：3759-3772
- 作者机构：
哈尔滨工业大学电子与信息工程学院，黑龙江哈尔滨 150001
- 作者简介：
[ "王浩添 男，2000年生，河北沧州人.现为哈尔滨工业大学硕士研究生.主要研究方向为SAR图像处理. E-mail: wanghaotian_hit@163.com" ]
[ "冀振元 男，1972年生，黑龙江五常人.现为哈尔滨工业大学教授、硕士生导师.主要研究方向为海态遥感、雷达成像、目标识别." ]
[ "化青龙 男，1995年生，安徽阜阳人.现为哈尔滨工业大学副研究员.主要研究方向为雷达图像处理和深度学习网络框架. E-mail: huaqinglong@hit.edu.cn" ]
[ "郭昭欣 男，2001年生，河南洛阳人.现为哈尔滨工业大学博士研究生.主要研究方向为雷达信号分析与智能信号处理.中国电子学会会员编号：E190185791A. E-mail: 1805919616@qq.com" ]
[ "张 云 女，1975年生，黑龙江虎林人.现为哈尔滨工业大学教授、博士生导师.主要研究方向为雷达信号处理、SAR成像、机器学习和遥感模式分析. E-mail: zhangyunhit@hit.edu.cn" ]
- 基金信息：国家自然科学基金(62401183);中国博士后科学基金(2024M754185;GZB20240967);黑龙江省博士后面上资助项目(LBH-Z24138)
DOI：10.12263/DZXB.20250356
中图分类号： TP751;收稿：2025-05-05，
修回：2025-10-12，
纸质出版：2025-10-25
- 稿件说明：
移动端阅览
王浩添, 冀振元, 化青龙, 等. 基于多分支多信息多深度复值特征融合网络的SAR舰船目标识别方法[J]. 电子学报, 2025, 53(10): 3759-3772.
WANG Hao-tian, JI Zhen-yuan, HUA Qing-long, et al. Recognition Method of Ship Targets for SAR Based on M3Net[J]. Acta Electronica Sinica, 2025, 53(10): 3759-3772.
王浩添, 冀振元, 化青龙, 等. 基于多分支多信息多深度复值特征融合网络的SAR舰船目标识别方法[J]. 电子学报, 2025, 53(10): 3759-3772. DOI：10.12263/DZXB.20250356
WANG Hao-tian, JI Zhen-yuan, HUA Qing-long, et al. Recognition Method of Ship Targets for SAR Based on M3Net[J]. Acta Electronica Sinica, 2025, 53(10): 3759-3772. DOI：10.12263/DZXB.20250356
摘要
针对合成孔径雷达（Synthetic Aperture Radar，SAR）图像中舰船目标识别任务中存在的类内差异显著与类间相似性高的难题，本文提出一种基于多分支多信息多深度复值特征融合网络（Multi-Branch， Multi-Information， Multi-Depth feature fusion complex-valued Network，M3Net）的舰船目标识别方法.传统方法多依赖人工设计的幅度特征，未能充分利用SAR原始数据中固有的复数特性，忽略了相位信息及其与幅度的耦合关系，导致对舰船精细结构的表征能力不足，识别精度与模型泛化能力受限.本文通过深入分析舰船目标的非圆性和复信号峰度特征，发现两者能够有效表征舰船目标区别于海面背景的散射特性，揭示了复数域统计量对舰船散射特性的表征优势.在此基础上，本文设计了深度复值特征提取模块（Complex Feature Extraction Module，CFEM），通过复卷积运算提取幅相耦合特征，创新性地引入实虚交融激活函数（Cross-fusion of Real and Imaginary Activation，CRIA），通过双激活函数的交叉耦合机制实现非线性特征交互，增强了对复数特征的表征能力.进一步构建多分支多信息多深度融合网络M3Net，通过主干复数域卷积神经网络（Complex-Valued Convolutional Neural Network，CV-CNN）、预训练CFEM分支和实值特征分支的协同处理，结合复数域注意力机制实现异构特征的动态加权融合，自适应突出判别性强的特征通道.在重构OpenSARship数据集上的实验结果表明，所提方法较传统CV-CNN提升5.89%，极差值降低至6.82%，显著改善了类别均衡性.
Abstract
To address the challenges of significant intra-class variations and high inter-class similarities in ship target recognition within synthetic aperture radar (SAR) images
this paper proposes a novel recognition method based on a multi-branch
multi-information
multi-depth feature fusion complex-valued network (M3Net). Traditional methods predominantly rely on manually designed amplitude features
failing to fully exploit the inherent complex-valued nature of raw SAR data and neglecting the crucial phase information and its coupling relationship with amplitude. This limitation results in insufficient characterization of ships’ fine structures and ultimately restricts recognition accuracy and model generalization capability. Through in-depth analysis of the noncircularity and complex signal kurtosis characteristics of ship targets
this study reveals that these features can effectively characterize the scattering properties distinguishing ships from the sea background
highlighting the representational advantages of complex-domain statistics for ship scattering characteristics. Building on this foundation
a deep complex feature extraction module (CFEM) is designed. This module employs complex-valued convolutional operations to extract amplitude-phase coupled features and innovatively introduces a cross-fusion of real and imaginary activation (CRIA) function. The CRIA mechanism
utilizing a dual-activation function cross-coupling approach
achieves nonlinear feature interactions and enhances the representational capacity for complex-valued features. Furthermore
the multi-branch
multi-information
multi-depth fusion network M3Net is constructed. M3Net synergistically integrates a core complex-valued convolutional neural network (CV-CNN) backbone
a pre-trained CFEM branch
and a real-valued feature branch. By incorporating a complex-domain attention mechanism
M3Net achieves dynamic weighted fusion of these heterogeneous features
adaptively highlighting the most discriminative feature channels. Experimental results on the reconstructed OpenSARship dataset demonstrate the effectiveness of the proposed method. Compared to the traditional CV-CNN
our approach achieves a 5.89% improvement in overall accuracy and reduces the maximum accuracy deviation across classes to 6.82%
significantly enhancing category balance.
关键词
Keywords
references
ZHANG C , ZHANG X , GAO G L , et al . Development and application of ship detection and classification datasets: A review [J ] . IEEE Geoscience and Remote Sensing Magazine , 2024 , 12 ( 4 ): 12 - 45 .
GAO G , WANG M X , ZHOU P , et al . A multibranch embedding network with bi-classifier for few-shot ship classification of SAR images [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2025 , 63 : 5201515 .
ZHANG Y P , LU D D , QIU X L , et al . Scattering-point topology for few-shot ship classification in SAR images [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2023 , 16 : 10326 - 10343 .
LI Y X , LAI X D , WANG M W , et al . C-SASO: A clustering-based size-adaptive safer oversampling technique for imbalanced SAR ship classification [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2022 , 60 : 5231112 .
RAJ J A , IDICULA S M , PAUL B . One-shot learning-based SAR ship classification using new hybrid siamese network [J ] . IEEE Geoscience and Remote Sensing Letters , 2022 , 19 : 4017205 .
WANG L , QI Y H , MATHIOPOULOS P T , et al . An improved SAR ship classification method using text-to-image generation-based data augmentation and squeeze and excitation [J ] . Remote Sensing , 2024 , 16 ( 7 ): 1299 .
LI L , WANG C , ZHANG H , et al . SAR image ship object generation and classification with improved residual conditional generative adversarial network [J ] . IEEE Geoscience and Remote Sensing Letters , 2022 , 19 : 4000105 .
XU Y J , LANG H T . Distribution shift metric learning for fine-grained ship classification in SAR images [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2020 , 13 : 2276 - 2285 .
HE J L , WANG Y H , LIU H W . Ship classification in medium-resolution SAR images via densely connected triplet cnns integrating fisher discrimination regularized metric learning [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2021 , 59 ( 4 ): 3022 - 3039 .
HUANG Z L , DATCU M , PAN Z X , et al . Deep SAR-Net: Learning objects from signals [J ] . ISPRS Journal of Photogrammetry and Remote Sensing , 2020 , 161 : 179 - 193 .
YU L J , HU Y H , XIE X C , et al . Complex-valued full convolutional neural network for SAR target classification [J ] . IEEE Geoscience and Remote Sensing Letters , 2020 , 17 ( 10 ): 1752 - 1756 .
ZHANG Y , YUAN H X , LI H B , et al . Complex-valued graph neural network on space target classification for defocused ISAR images [J ] . IEEE Geoscience and Remote Sensing Letters , 2022 , 19 : 4512905 .
ZHOU X Q , LUO C , REN P , et al . Multiscale complex-valued feature attention convolutional neural network for SAR automatic target recognition [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2024 , 17 : 2052 - 2066 .
ZENG Z Q , SUN J P , HAN Z , et al . SAR automatic target recognition method based on multi-stream complex-valued networks [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2022 , 60 : 5228618 .
WANG R N , WANG Z C , XIA K W , et al . Target recognition in single-channel SAR images based on the complex-valued convolutional neural network with data augmentation [J ] . IEEE Transactions on Aerospace and Electronic Systems , 2023 , 59 ( 2 ): 796 - 804 .
方澄 , 管方恒 , 李天驰 , 等 . 端到端全复数域SAR目标分类神经网络 [J ] . 电子学报 , 2024 , 52 ( 7 ): 2449 - 2460 .
FANG C , GUAN F H , LI T C , et al . End-to-end full complex-valued domain SAR target classification neural network [J ] . Acta Electronica Sinica , 2024 , 52 ( 7 ): 2449 - 2460 . (in Chinese)
HAO D D , ZHANG Z , LU D D , et al . CVGG-Net: Ship recognition for SAR images based on complex-valued convolutional neural network [J ] . IEEE Geoscience and Remote Sensing Letters , 2023 , 20 : 4010805 .
ZHANG Y , HUA Q L , JIANG Y C , et al . CV-MotionNet: Complex-valued convolutional neural network for SAR moving ship targets classification [C ] // 2021 IEEE International Geoscience and Remote Sensing Symposium IGARSS . Piscataway : IEEE , 2021 : 4280 - 4283 .
HUA Q L , ZHANG Y , WEI C X , et al . CV-RotNet: Complex-valued convolutional neural network for SAR three-dimensional rotating ship target recognition [C ] // IGARSS 2022-2022 IEEE International Geoscience and Remote Sensing Symposium . Piscataway : IEEE , 2022 : 3552 - 3555 .
张云 , 化青龙 , 姜义成 , 等 . 基于混合型复数域卷积神经网络的三维转动舰船目标识别 [J ] . 电子学报 , 2022 , 50 ( 5 ): 1042 - 1049 .
ZHANG Y , HUA Q L , JIANG Y C , et al . Recognition of 3D rotating ship based on Mix-CV-CNN [J ] . Acta Electronica Sinica , 2022 , 50 ( 5 ): 1042 - 1049 . (in Chinese)
WANG H T , JI Z Y , HUA Q L , et al . CRIA: An enhancement method for CV-CNN based on cross-fusion of complex information of real and imaginary activations [C ] // 2024 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) . Athens : IEEE , 2024 : 10285 - 10288 .
LENG X G , JI K F , ZHOU S L , et al . Noncircularity parameters and their potential in ship detection from high resolution SAR imagery [C ] // 2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS) . Piscataway : IEEE , 2017 : 1876 - 1879 .
BRADLEY D , MORRIS J M , ADALI T , et al . On the detection of RFI using the complex signal kurtosis in microwave radiometry [C ] // 2014 13th Specialist Meeting on Microwave Radiometry and Remote Sensing of the Environment (MicroRad) . Piscataway : IEEE , 2014 : 33 - 38 .
冷祥光 . 基于单通道SAR图像复值信息的舰船目标检测技术研究 [D ] . 长沙 : 国防科技大学 , 2019 .
LENG X G . Study on Ship Detection Technology Based on Complex-Valued Information in Single-Channel SAR Imagery [D ] . Changsha : National University of Defense Technology , 2019 . (in Chinese)
LENG X G , JI K F , ZHOU S L , et al . Ship detection based on complex signal kurtosis in single-channel SAR imagery [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2019 , 57 ( 9 ): 6447 - 6461 .
HUANG L Q , LIU B , LI B Y , et al . OpenSARShip: A dataset dedicated to Sentinel-1 ship interpretation [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2018 , 11 ( 1 ): 195 - 208 .
ZHANG T W , ZHANG X L . Squeeze-and-excitation laplacian pyramid network with dual-polarization feature fusion for ship classification in SAR images [J ] . IEEE Geoscience and Remote Sensing Letters , 2022 , 19 : 4019905 .
MAKHIJA S , MANDAL S , PANDYA U , et al . PolSAR image classification using complex-valued squeeze and excitation networ [M ] // Pattern Recognition . Cham : Springer Nature Switzerland , 2024 : 270 - 286 .
AMARI S I . Backpropagation and stochastic gradient descent method [J ] . Neurocomputing , 1993 , 5 ( 4/5 ): 185 - 196 .
ZHANG Y P , ZHANG Q , KANG L , et al . End-to-end recognition of similar space cone-cylinder targets based on complex-valued coordinate attention networks [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2021 , 60 : 5106214 .
PAN H X , YANG C , YIN Y F , et al . Feature attention complex-valued-based CNN for ship target recognition of SAR images [C ] // 2024 IEEE International Conference on Signal, Information and Data Processing . Piscataway : IEEE , 2025 : 1 - 5 .
ZHANG T W , ZHANG X L , KE X , et al . HOG-ShipCLSNet: A novel deep learning network with hog feature fusion for SAR ship classification [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2022 , 60 : 5210322 .
0
浏览量
7
下载量
0
CSCD
- 网络PDF下载
- Meta-XML下载
- BibTeX下载
- Endnote下载
- RefMan 下载
- NoteExpress下载
- NoteFirst下载
关联资源
相关文章
相关作者
相关机构
- 技术支持由北京北大方正电子有限公司提供 京ICP备09064830号-19
京公网安备11010802024621 - 本系统建议在Chrome、 IE9+ 以上版本浏览器阅读本站内容，360浏览器请切换至极速模式
- Cookies帮助我们提供服务并提供个性化体验。使用本网站，即表示您同意我们使用Cookies