基于频域感知与原型对齐的多源域泛化SAR舰船检测方法

艾加秋; 朱雅喃; 郭丹; 庄硕; 陈立福

doi:10.12263/DZXB.20260607

您当前的位置：

首页 >

文章列表页 >

基于频域感知与原型对齐的多源域泛化SAR舰船检测方法

更新时间：2026-08-06

- 基于频域感知与原型对齐的多源域泛化SAR舰船检测方法
- Multi-Source Domain Generalization for SAR Ship Detection Based on Frequency-Aware and Prototype Alignment
- 电子学报 2026年页码：1-12
- 作者机构：
  
  1.合肥工业大学计算机与信息学院，安徽合肥 230009
  2.安徽省智能互联系统实验室（合肥工业大学），安徽合肥 230009
  3.长沙理工大学人工智能学院，湖南长沙 410114
- 作者简介：
  
  艾加秋男，1985年8月出生于江西省吉安市。现为合肥工业大学计算机与信息学院教授、博士生导师。主要研究方向为多模态遥感图像智能处理与应用。E-mail: aijiaqiu1985@hfut.edu.cn
  朱雅喃女，1997年12月出生于山东省青岛市。现为合肥工业大学计算机科学与技术流动站博士后。主要研究方向为域适应、跨模态遥感图像智能解译。E-mail: 2026820020@hfut.edu.cn
- 基金信息：
  
  国家自然科学基金面上项目(62471171;62071164);合肥市揭榜挂帅重大项目(2024SZD004);中央高校基本科研业务费专项资金(PA2025IISL0111)
- DOI：10.12263/DZXB.20260607
  中图分类号： TN958;TP391.41
- 收稿：2026-06-15，
  
  录用：2026-07-02，
  
  网络首发：2026-08-06，
- 稿件说明：
移动端阅览
艾加秋, 朱雅喃, 郭丹, 等. 基于频域感知与原型对齐的多源域泛化SAR舰船检测方法[J/OL]. 电子学报, 2026,1-12.

AI Jiaqiu, ZHU Yanan, GUO Dan, et al. Multi-Source Domain Generalization for SAR Ship Detection Based on Frequency-Aware and Prototype Alignment[J/OL]. ACTA ELECTRONICA SINICA, 2026, 1-12.
艾加秋, 朱雅喃, 郭丹, 等. 基于频域感知与原型对齐的多源域泛化SAR舰船检测方法[J/OL]. 电子学报, 2026,1-12. DOI： 10.12263/DZXB.20260607.

AI Jiaqiu, ZHU Yanan, GUO Dan, et al. Multi-Source Domain Generalization for SAR Ship Detection Based on Frequency-Aware and Prototype Alignment[J/OL]. ACTA ELECTRONICA SINICA, 2026, 1-12. DOI： 10.12263/DZXB.20260607.

摘要

合成孔径雷达（Synthetic Aperture Radar，SAR）是一种主动微波遥感成像技术，具备全天时、全天候的工作能力，在海洋环境感知中具有不可替代的作用。SAR舰船检测在海洋监管、海上安全保障和海上交通管理等领域发挥着重要作用，是海上目标感知和安全检测中的重要技术手段，尤其是在复杂海洋环境下。但入射角、海况和噪声水平会导致成像条件不同，SAR舰船数据存在明显的数据分布差异，导致模型在未知场景下不能达到稳定的舰船检测性能。域泛化技术是指仅利用源域数据来学习域不变特征表示，进而提升模型在未见目标域中的舰船检测能力。多源域泛化为学习鲁棒的舰船表征提供了必要的条件，但同时也加大了域间偏移，提升了优化难度。针对上述问题，本文提出一种基于频域感知与原型对齐的多源域泛化SAR舰船检测方法。该方法从频域建模、语义约束和源域加权优化这三个方面联合优化，以综合提高模型的跨域鲁棒性能。具体而言，由于SAR图像上的斑点噪声、海杂波与舰船结构在频域中有不同表现，本文设计了频域感知特征增强模块，利用傅里叶变换对特征进行分解，然后将低频结构信息和高频噪声成分重新加权，从而增强与舰船结构有关的域不变特征，抑制域相关扰动。其次，为缓解不同源域中的舰船目标由成像条件所导致的语义偏移，本文基于候选特征来构建各个源域的舰船语义原型，然后在特征空间里进行显式的对齐，进而达到跨域一致的语义表示，以减少舰船目标之间的语义偏差。最后，考虑到各个源域给模型带来的贡献大小是不一样的，部分源域与未知目标域的分布更加接近，但是部分源域由于噪声过大、样本质量差等原因会对模型造成影响，本文提出了多源自适应加权策略。该策略会根据训练损失来动态分配权重，进而有效地降低不可靠源域所带来的负迁移影响。实验结果表明，本文所提方法在多个SAR舰船检测数据集上显著优于现有方法，且舰船检测精度和泛化能力均取得了显著提升。

Abstract

Synthetic Aperture Radar (SAR) is an active microwave remote sensing technique with all-day and all-weather imaging capabilities

playing an irreplaceable role in maritime environment perception. SAR ship detection is essential for target perception and maritime safety monitoring in complex marine environments. However

variations in imaging conditions

including incidence angle

sea state

and noise level

often cause substantial distribution shifts in SAR data

leading to unstable detection performance in unseen scenarios. Domain generalization addresses this issue by learning domain-invariant representations from source domains to improve detection in unseen target domains. In multi-source settings

diverse source distributions offer richer supervision for robust ship representation learning

but also introduce more complex inter-domain discrepancies and optimization challenges. To tackle these challenges

we propose a frequency-aware and prototype-aligned multi-source domain generalization framework for SAR ship detection. The framework enhances cross-domain robustness from three aspects. Specifically

considering that speckle noise

sea clutter

and ship structures exhibit different characteristics in the frequency domain

a frequency-aware feature enhancement module is designed. This module decomposes features using Fourier transform and reweights low-frequency structural information and high-frequency noise components

thereby enhancing domain-invariant ship-related features and suppressing domain-specific disturbances. Moreover

to alleviate semantic shifts caused by different imaging conditions across source domains

ship semantic prototypes are constructed from candidate features and explicitly aligned in the feature space

leading to more consistent cross-domain semantic representations. Finally

considering that different source domains contribute unequally to model training and that some domains may negatively affect optimization due to severe noise or poor sample quality

a multi-source adaptive weighting strategy is introduced. This strategy dynamically assigns weights according to training losses

effectively reducing negative transfer caused by unreliable source domains. Experiments on multiple SAR ship detection datasets show that the proposed method consistently outperforms existing approaches in cross-domain scenarios

achieving improved detection accuracy and generalization capability.

关键词

Keywords

references

胡艳芬 , 王宝发 . 多极化SAR图像的降斑研究 [J ] . 电子学报 , 2003 , 31 ( 2 ): 207 - 210 . DOI: 10.3321/j.issn:0372-2112.2003.02.013 http://dx.doi.org/10.3321/j.issn:0372-2112.2003.02.013

Hu Yanfen , Wang Baofa . Speckle reduction in multipolarization SAR imagery [J ] . Acta Electronica Sinica , 2003 , 31 ( 2 ): 207 - 210 . (in Chinese) . DOI: 10.3321/j.issn:0372-2112.2003.02.013 http://dx.doi.org/10.3321/j.issn:0372-2112.2003.02.013

鲍厚兵 , 黄顺吉 , 刘钊 . SAR并行成像处理的研究 [J ] . 电子学报 , 2002 , 30 ( 12 ): 1745 - 1747 .

Bao Houbing , Huang Shunji , Liu Zhao . The study of parallel processing of SAR imaging [J ] . Acta Electronica Sinica , 2002 , 30 ( 12 ): 1745 - 1747 . (in Chinese)

Li Dong , Liang Quanhuan , Liu Hongqing , et al . A novel multidimensional domain deep learning network for SAR ship detection [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2022 , 60 : 5203213 . DOI: 10.1109/tgrs.2021.3062038 http://dx.doi.org/10.1109/tgrs.2021.3062038

Chen Yulin , Shen Yanyun , Duan Chi , et al . Robust and efficient SAR ship detection: An integrated despecking and detection framework [J ] . Remote Sensing , 2025 , 17 ( 4 ): 580 . DOI: 10.3390/rs17040580 http://dx.doi.org/10.3390/rs17040580

Guan Tianyue , Chang Sheng , Wang Chunle , et al . SAR small ship detection based on enhanced YOLO network [J ] . Remote Sensing , 2025 , 17 ( 5 ): 839 . DOI: 10.3390/rs17050839 http://dx.doi.org/10.3390/rs17050839

Xue Weibao , Ai Jiaqiu , Zhu Yanan , et al . LMCNet: Lightweight modality compensation network via knowledge distillation for salient ship detection under missing-modality conditions [J ] . IEEE Transactions on Aerospace and Electronic Systems , 2026 , 62 : 6547 - 6560 . DOI: 10.1109/taes.2026.3664356 http://dx.doi.org/10.1109/taes.2026.3664356

Ren Haohao , Liang Shuyi , Miao Lei , et al . Multi-granularity feature fusion network integrating ASC for SAR target recognition [J ] . IEEE Geoscience and Remote Sensing Letters , 2026 , 23 : 4003805 . DOI: 10.1109/lgrs.2025.3649343 http://dx.doi.org/10.1109/lgrs.2025.3649343

Zhou Yun , Zhu Yinglin , Ren Haohao , et al . Refined multi-modal feature learning framework for marine target detection using radar sensor [J ] . Digital Signal Processing , 2026 , 170 : 105816 . DOI: 10.1016/j.dsp.2025.105816 http://dx.doi.org/10.1016/j.dsp.2025.105816

Ai Jiaqiu , Tian Ruitian , Luo Qiwu , et al . Multi-scale rotation-invariant Haar-like feature integrated CNN-based ship detection algorithm of multiple-target environment in SAR imagery [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2019 , 57 ( 12 ): 10070 - 10087 . DOI: 10.1109/tgrs.2019.2931308 http://dx.doi.org/10.1109/tgrs.2019.2931308

Shi Yu , Du Lan , Guo Yuchen . Unsupervised domain adaptation for SAR target detection [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2021 , 14 : 6372 - 6385 . DOI: 10.1109/jstars.2021.3089238 http://dx.doi.org/10.1109/jstars.2021.3089238

Huang Jinyang , Feng Yuanhao , Cui Fengqi , et al . Identifying who you are no matter what you write through abstracting handwriting style [J ] . IEEE Transactions on Dependable and Secure Computing , 2026 , 23 ( 3 ): 6890 - 6905 . DOI: 10.1109/tdsc.2026.3668275 http://dx.doi.org/10.1109/tdsc.2026.3668275

Li Da , Yang Yongxin , Song Yizhe , et al . Deeper, broader and artier domain generalization [C ] // Proceedings of the IEEE International Conference on Computer Vision . Piscataway : IEEE , 2017 : 5542 - 5550 . DOI: 10.1109/iccv.2017.591 http://dx.doi.org/10.1109/iccv.2017.591

Zhong Yingyi , Zhou Wen’an , Wang Zhixian . A survey of data augmentation in domain generalization [J ] . Neural Processing Letters , 2025 , 57 ( 2 ): 34 . DOI: 10.1007/s11063-025-11747-9 http://dx.doi.org/10.1007/s11063-025-11747-9

Zhang Zenghui , Zhang Limeng , Wu Juanping , et al . Optical and synthetic aperture radar image fusion for ship detection and recognition: Current state, challenges, and future prospects [J ] . IEEE Geoscience and Remote Sensing Magazine , 2024 , 12 ( 4 ): 132 - 168 . DOI: 10.1109/mgrs.2024.3404506 http://dx.doi.org/10.1109/mgrs.2024.3404506

Han Jianhong , Yang Wenjie , Wang Yupei , et al . Remote sensing teacher: Cross-domain detection transformer with learnable frequency-enhanced feature alignment in remote sensing imagery [J ] . IEEE Transactions on Geoscience and Remote Sensing , 2024 , 62 : 5619814 . DOI: 10.1109/tgrs.2024.3378284 http://dx.doi.org/10.1109/tgrs.2024.3378284

Ren Shaoqing , He Kaiming , Girshick R , et al . Faster R-CNN: Towards real-time object detection with region proposal networks [J ] . IEEE Transactions on Pattern Analysis and Machine Intelligence , 2017 , 39 ( 6 ): 1137 - 1149 . DOI: 10.1109/tpami.2016.2577031 http://dx.doi.org/10.1109/tpami.2016.2577031

Redmon J , Divvala S , Girshick R , et al . You only look once: Unified, real-time object detection [C ] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2016 : 779 - 788 . DOI: 10.1109/cvpr.2016.91 http://dx.doi.org/10.1109/cvpr.2016.91

Lin T Y , Dollár P , Girshick R , et al . Feature pyramid networks for object detection [C ] // Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2017 : 936 - 944 . DOI: 10.1109/cvpr.2017.106 http://dx.doi.org/10.1109/cvpr.2017.106

Cao Rui , Sui Jianghua . A dynamic multi-scale feature fusion network for enhanced SAR ship detection [J ] . Sensors , 2025 , 25 ( 16 ): 5194 . DOI: 10.3390/s25165194 http://dx.doi.org/10.3390/s25165194

Ding Jian , Xue Nan , Long Yang , et al . Learning RoI transformer for oriented object detection in aerial images [C ] // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2019 : 2844 - 2853 . DOI: 10.1109/cvpr.2019.00296 http://dx.doi.org/10.1109/cvpr.2019.00296

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 Haotian , Ji Zhenyuan , Hua Qinglong , et al . Recognition method of ship targets for SAR based on M3Net [J ] . Acta Electronica Sinica , 2025 , 53 ( 10 ): 3759 - 3772 . (in Chinese)

Shan Huilin , Fu Xiangwei , Lv Zongkui , et al . SAR ship detection algorithm based on deep dense sim attention mechanism network [J ] . IEEE Sensors Journal , 2023 , 23 ( 14 ): 16032 - 16041 . DOI: 10.1109/jsen.2023.3284959 http://dx.doi.org/10.1109/jsen.2023.3284959

Zhu Yanan , Ai Jiaqiu , Xue Weibao , et al . Cross-modal ship detection from optical to SAR images based on pixel- and feature-level progressive transfer [J ] . IEEE Sensors Journal , 2025 , 25 ( 8 ): 13344 - 13356 . DOI: 10.1109/jsen.2025.3543520 http://dx.doi.org/10.1109/jsen.2025.3543520

Kim J U , Park S , Ro Y M . Uncertainty-guided cross-modal learning for robust multispectral pedestrian detection [J ] . IEEE Transactions on Circuits and Systems for Video Technology , 2022 , 32 ( 3 ): 1510 - 1523 . DOI: 10.1109/tcsvt.2021.3076466 http://dx.doi.org/10.1109/tcsvt.2021.3076466

Chen Xi , Wang Zhirui , Wang Wenhao , et al . CroMoDa: Unsupervised oriented SAR ship detection via cross-modality distribution alignment [J ] . IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2024 , 17 : 11899 - 11914 . DOI: 10.1109/jstars.2024.3420901 http://dx.doi.org/10.1109/jstars.2024.3420901

Lin Mingkai , Li Wenzhong , Li Ding , et al . Multi-domain generalized graph meta learning [C ] // Proceedings of the AAAI Conference on Artificial Intelligence . Palo Alto : AAAI , 2023 : 4479 - 4487 . DOI: 10.1609/aaai.v37i4.25569 http://dx.doi.org/10.1609/aaai.v37i4.25569

Liu Hong , Song Pinhao , Ding Runwei . Towards domain generalization in underwater object detection [C ] // Proceedings of 2020 IEEE International Conference on Image Processing (ICIP) . Piscataway : IEEE , 2020 : 1971 - 1975 . DOI: 10.1109/icip40778.2020.9191364 http://dx.doi.org/10.1109/icip40778.2020.9191364

Addepalli S , Asokan A R , Sharma L , et al . Leveraging vision-language models for improving domain generalization in image classification [C ] // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2024 : 23922 - 23932 . DOI: 10.1109/cvpr52733.2024.02258 http://dx.doi.org/10.1109/cvpr52733.2024.02258

Vidit V , Engilberge M , Salzmann M . CLIP the gap: A single domain generalization approach for object detection [C ] // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2023 : 3219 - 3229 . DOI: 10.1109/cvpr52729.2023.00314 http://dx.doi.org/10.1109/cvpr52729.2023.00314

Danish M S , Khan M H , Munir M A , et al . Improving single domain-generalized object detection: A focus on diversification and alignment [C ] // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2024 : 17732 - 17742 . DOI: 10.1109/cvpr52733.2024.01679 http://dx.doi.org/10.1109/cvpr52733.2024.01679

Al-Emadi S A , Yang Y , Ofli F . Benchmarking object detectors under real-world distribution shifts in satellite imagery [C ] // Proceedings of the Computer Vision and Pattern Recognition Conference . Piscataway : IEEE , 2025 : 8299 - 8309 . DOI: 10.1109/cvpr52734.2025.00777 http://dx.doi.org/10.1109/cvpr52734.2025.00777

Dou Qi , Castro D C , Kamnitsas K , et al . Domain generalization via model-agnostic learning of semantic features [C ] // Proceedings of the 33rd International Conference on Neural Information Processing Systems . New York : Curran Associates Inc. , 2019 : 579 .

He Boyong , Ji Yuxiang , Ye Qianwen , et al . Generalized diffusion detector: Mining robust features from diffusion models for domain-generalized detection [C ] // Proceedings of the Computer Vision and Pattern Recognition Conference . Piscataway : IEEE , 2025 : 9921 - 9932 . DOI: 10.1109/cvpr52734.2025.00927 http://dx.doi.org/10.1109/cvpr52734.2025.00927

Wu Aming , Deng Cheng . Single-domain generalized object detection in urban scene via cyclic-disentangled self-distillation [C ] // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2022 : 837 - 846 . DOI: 10.1109/cvpr52688.2022.00092 http://dx.doi.org/10.1109/cvpr52688.2022.00092

Lee S , Seong H , Lee S , et al . WildNet: Learning domain generalized semantic segmentation from the wild [C ] // Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2022 : 9936 - 9946 . DOI: 10.1109/cvpr52688.2022.00970 http://dx.doi.org/10.1109/cvpr52688.2022.00970

Duraklı E , Aptoula E . Domain generalized object detection for remote sensing images [C ] // Proceedings of 2023 31st Signal Processing and Communications Applications Conference (SIU) . Piscataway : IEEE , 2023 : 1 - 4 . DOI: 10.1109/siu59756.2023.10223771 http://dx.doi.org/10.1109/siu59756.2023.10223771

Lee W , Hong D , Lim H , et al . Object-aware domain generalization for object detection [C ] // Proceedings of the AAAI Conference on Artificial Intelligence . Palo Alto : AAAI , 2024 : 2947 - 2955 . DOI: 10.1609/aaai.v38i4.28076 http://dx.doi.org/10.1609/aaai.v38i4.28076

Wu Fan , Gao Jinling , Hong Lanqing , et al . G-NAS: Generalizable neural architecture search for single domain generalization object detection [C ] // Proceedings of the AAAI Conference on Artificial Intelligence . Palo Alto : AAAI , 2024 : 5958 - 5966 . DOI: 10.1609/aaai.v38i6.28410 http://dx.doi.org/10.1609/aaai.v38i6.28410

Chen Yang , Song Pinhao , Liu Hong , et al . Achieving domain generalization for underwater object detection by domain mixup and contrastive learning [J ] . Neurocomputing , 2023 , 528 : 20 - 34 . DOI: 10.1016/j.neucom.2023.01.053 http://dx.doi.org/10.1016/j.neucom.2023.01.053

Yao Xingxu , Zhao Sicheng , Xu Pengfei , et al . Multi-source domain adaptation for object detection [C ] // Proceedings of the IEEE/CVF International Conference on Computer Vision . Piscataway : IEEE , 2021 : 3273 - 3282 . DOI: 10.1109/iccv48922.2021.00326 http://dx.doi.org/10.1109/iccv48922.2021.00326

Belal A , Meethal A , Romero F P , et al . Multi-source domain adaptation for object detection with prototype-based mean teacher [C ] // Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision . Piscataway : IEEE , 2024 : 1266 - 1275 . DOI: 10.1109/wacv57701.2024.00131 http://dx.doi.org/10.1109/wacv57701.2024.00131

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

多环境选择策略引导的约束多目标进化优化算法

基于动态图神经网络与对比学习的单通道同频盲信号分离方法

中继收发器未知的TDOA/TDOA-AOA无源定位算法

基于DPU的机载射频综合处理系统研究与设计

金刚石高压场效应管单粒子效应研究