Few-Shot SAR Target Recognition Method Based on Feature Fusion Attention and Meta-ResNet

LIU Qi; LIU Yong-xiang; ZHANG Xin-yu

doi:10.12263/DZXB.20211371

您当前的位置：

首页 >

文章列表页 >

Few-Shot SAR Target Recognition Method Based on Feature Fusion Attention and Meta-ResNet

PAPERS | 更新时间：2025-12-08

- Few-Shot SAR Target Recognition Method Based on Feature Fusion Attention and Meta-ResNet
- ACTA ELECTRONICA SINICA Vol. 51, Issue 9, Pages: 2366-2378(2023)
- 作者机构：
  
  国防科技大学电子科学学院，湖南长沙 410073
- 作者简介：
- 基金信息：
  
  National Natural Science Foundation of China(61025006;60872134;61901482;61921001);China Postdoctoral Science Foundation(2018M633667)
- DOI：10.12263/DZXB.20211371
  CLC： TN95;TN911.7
- Received：11 October 2021，
  
  Revised：2022-07-10，
  
  Published：25 September 2023
- 稿件说明：
移动端阅览
刘旗,刘永祥,张新禹.基于特征注意力融合元残差网络的小样本SAR目标识别[J].电子学报,2023,51(09):2366-2378.

LIU Qi,LIU Yong-xiang,ZHANG Xin-yu.Few-Shot SAR Target Recognition Method Based on Feature Fusion Attention and Meta-ResNet[J].ACTA ELECTRONICA SINICA,2023,51(09):2366-2378.
刘旗,刘永祥,张新禹.基于特征注意力融合元残差网络的小样本SAR目标识别[J].电子学报,2023,51(09):2366-2378. DOI： 10.12263/DZXB.20211371.

LIU Qi,LIU Yong-xiang,ZHANG Xin-yu.Few-Shot SAR Target Recognition Method Based on Feature Fusion Attention and Meta-ResNet[J].ACTA ELECTRONICA SINICA,2023,51(09):2366-2378. DOI： 10.12263/DZXB.20211371.

摘要

深度学习方法在合成孔径雷达（Synthetic Aperture Radar， SAR）目标识别方面表现良好，但是这些深度模型往往需要大量训练数据来对模型参数进行优化，否则经常会遇到严重的过拟合问题，导致目标识别准确率低，模型泛化能力差.针对上述问题，基于元学习框架，提出了针对小样本条件下SAR目标识别的特征注意力融合元残差网络.在该方法中，基于残差网络设计了全新的学习器，通过利用残差结构，有效传递SAR图像的对比度信息，提高目标识别准确率，而且多尺度特征注意力融合模块通过注意力机制，融合不同残差层特征，为目标识别提供更丰富的特征信息.改进的元学习器不仅能够为学习器学习到易于优化的初始化参数，而且能够为学习器的每一个参数学习一个不同但是合适的学习率.与其他三种小样本目标识别方法在MSTAR数据集上进行对比实验，结果表明本文方法提高了小样本条件下SAR目标识别方法的识别准确率和识别模型的泛化性能.对所提方法进行了鲁棒性验证实验，网络结构消融实验，并展示了元学习器为学习器参数学习到的不同但是合适的学习率.

Abstract

In recent years

traditional synthetic aperture radar (SAR) target recognition methods based on deep learning have achieved promising results. However

these deep models need lots of training samples for parameter optimization

otherwise they would possibly encounter serious overfitting problem

resulting in low recognition accuracy and poor generalization ability of these deep models. Aiming to above problems

this paper proposed a few-shot SAR target recognition method named multi feature attention fusion Meta-ResNet. In this method

we designed a learner based on residual network

which can effectively transmit contrast information in SAR images and thus improve recognition accuracy. Furthermore

the multi feature attention fusion module can provide weighted multi scale features by fusing features of different ResNet layers. The improved meta-learner can not only learn good initialization parameters for learner

but also learn a different but appropriate learning rate for each learner parameter. The comparative experiment between the proposed method and other three few-shot recognition methods demonstrated the effectiveness and progressiveness of the proposed method. We also conducted experiment to verify the robustness and study the influence of the network structure on recognition accuracy. We showed the different but appropriate learning rates of learner parameters learned by the meta-learner.

关键词

Keywords

references

BARTON D K , LEONOV S A . Radar Technology Encyclopedia [M]. London : Artech House , 1998 .

胡利平 . 合成孔径雷达图像目标识别技术研究 [D]. 西安 : 西安电子科技大学 , 2009 .

HU L P . Study on SAR Images Target Recognition [D]. Xi'an : Xidian University , 2009 . (in Chinese)

庞礴 . 多通道SAR成像理论与方法研究 [D]. 长沙 : 国防科学技术大学 , 2014 .

PANG B . Study on the Imaging Theory and Method of Multi-Channel SAR [D]. Changsha : National University of Defense Technology , 2014 . (in Chinese)

LI Y Y , ZHOU L H , PENG C , et al . Spatial fuzzy clustering and deep auto-encoder for unsupervised change detection in synthetic aperture radar images [C]// IGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium . Piscataway : IEEE , 2018 : 4479 - 4482 .

RAN L , LIU Z , LI T , et al . Extension of map-drift algorithm for highly squinted SAR autofocus [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2017 , 10 ( 9 ): 4032 - 4044 .

SHI L J , KARVONEN J , CHENG B , et al . Sea ice thickness retrieval from SAR imagery over Bohai Sea [C]// 2014 IEEE Geoscience and Remote Sensing Symposium . Piscataway : IEEE , 2014 : 4864 - 4867 .

赵鹏菲 , 黄丽佳 . 一种基于EfficientNet与BiGRU的多角度SAR图像目标识别方法 [J]. 雷达学报 , 2021 , 10 ( 6 ): 895 - 904 .

ZHAO P F , HUANG L J . Target recognition method for multi-aspect synthetic aperture radar images based on EfficientNet and BiGRU [J]. Journal of Radars , 2021 , 10 ( 6 ): 895 - 904 . (in Chinese)

郭倩 , 王海鹏 , 徐丰 . SAR图像飞机目标检测识别进展 [J]. 雷达学报 , 2020 , 9 ( 3 ): 497 - 513 .

GUO Q , WANG H P , XU F . Research progress on aircraft detection and recognition in SAR imagery [J]. Journal of Radars , 2020 , 9 ( 3 ): 497 - 513 . (in Chinese)

ZHANG J S , XING M D , XIE Y Y . FEC: A feature fusion framework for SAR target recognition based on electromagnetic scattering features and deep CNN features [J]. IEEE Transactions on Geoscience and Remote Sensing , 2021 , 59 ( 3 ): 2174 - 2187 .

CHEN L , JIANG X , LI Z , et al . Feature-enhanced speckle reduction via low-rank and space-angle continuity for circular SAR target recognition [J]. IEEE Transactions on Geoscience and Remote Sensing , 2020 , 58 ( 11 ): 7734 - 7752 .

OH J , YOUM G Y , KIM M . SPAM-net: A CNN-based SAR target recognition network with pose angle marginalization learning [J]. IEEE Transactions on Circuits and Systems for Video Technology , 2021 , 31 ( 2 ): 701 - 714 .

GUO Y C , DU L , WEI D , et al . Robust SAR automatic target recognition via adversarial learning [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2020 , 14 : 716 - 729 .

CHO J H , PARK C G . Multiple feature aggregation using convolutional neural networks for SAR image-based automatic target recognition [J]. IEEE Geoscience and Remote Sensing Letters , 2018 , 15 ( 12 ): 1882 - 1886 .

GAO F , HUANG T , SUN J P , et al . A new algorithm for SAR image target recognition based on an improved deep convolutional neural network [J]. Cognitive Computation , 2019 , 11 ( 6 ): 809 - 824 .

YANG R , HU Z T , LIU Y M , et al . A novel polarimetric SAR classification method integrating pixel-based and patch-based classification [J]. IEEE Geoscience and Remote Sensing Letters , 2020 , 17 ( 3 ): 431 - 435 .

LI L P , LIU J , SU L Y , et al . A novel graph metalearning method for SAR target recognition [J]. IEEE Geoscience and Remote Sensing Letters , 2022 , 19 : 1 - 5 .

MARGARIT G , MALLORQUI J J , FORTUNY-GUASCH J , et al . Exploitation of ship scattering in polarimetric SAR for an improved classification under high clutter conditions [J]. IEEE Transactions on Geoscience and Remote Sensing , 2009 , 47 ( 4 ): 1224 - 1235 .

张晰 , 张杰 , 纪永刚 , 等 . 基于结构特征的SAR船只类型识别能力分析 [J]. 海洋学报 , 2010 , 32 ( 1 ): 146 - 152 .

ZHANG X , ZHANG J , JI Y G , et al . The capability analysis of ship classification by structure feature using SAR images [J]. Acta Oceanologica Sinica , 2010 , 32 ( 1 ): 146 - 152 . (in Chinese)

RAEISI A , AKBARIZADEH G , MAHMOUDI A . Combined method of an efficient cuckoo search algorithm and nonnegative matrix factorization of different zernike moment features for discrimination between oil spills and lookalikes in SAR images [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2018 , 11 ( 11 ): 4193 - 4205 .

DING B Y , WEN G J . Target reconstruction based on 3-D scattering center model for robust SAR ATR [J]. IEEE Transactions on Geoscience and Remote Sensing , 2018 , 56 ( 7 ): 3772 - 3785 .

SUKAWATTANAVIJIT C , CHEN J , ZHANG H S . GA-SVM algorithm for improving land-cover classification using SAR and optical remote sensing data [J]. IEEE Geoscience and Remote Sensing Letters , 2017 , 14 ( 3 ): 284 - 288 .

SUN Y J , LIU Z P , TODOROVIC S , et al . Adaptive boosting for SAR automatic target recognition [J]. IEEE Transactions on Aerospace and Electronic Systems , 2007 , 43 ( 1 ): 112 - 125 .

SRINIVAS U , MONGA V , RAJ R G . SAR automatic target recognition using discriminative graphical models [J]. IEEE Transactions on Aerospace and Electronic Systems , 2014 , 50 ( 1 ): 591 - 606 .

HUA W Q , ZHANG C , XIE W , et al . Polarimetric SAR image classification based on ensemble dual-branch CNN and superpixel algorithm [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2022 , 5 : 2759 - 2772 .

ZHAO Z Q , JIAO L C , ZHAO J Q , et al . Discriminant deep belief network for high-resolution SAR image classification [J]. Pattern Recognition , 2017 , 61 : 686 - 701 .

WANG J L , HOU B , JIAO L C , et al . POL-SAR image classification based on modified stacked autoencoder network and data distribution [J]. IEEE Transactions on Geoscience and Remote Sensing , 2020 , 58 ( 3 ): 1678 - 1695 .

QIN F C , GUO J M , SUN W D . Object-oriented ensemble classification for polarimetric SAR imagery using restricted Boltzmann machines [J]. Remote Sensing Letters , 2017 , 8 ( 3 ): 204 - 213 .

HE K M , ZHANG X Y , REN S Q , et al . Deep residual learning for image recognition [C]// 2016 IEEE Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2016 : 770 - 778 .

VINYALS O , BLUNDELL C , LILLICRAP T , et al . Matching networks for one shot learning [C]// Proceedings of the 30th International Conference on Neural Information Processing Systems . 2016 : 3637 - 3645 .

RAVI S , LAROCHELLE H . Optimization as a model for few-shot learning [EB/OL]. ( 2016-11-05 )[ 2021-09-11 ]. https://openreview.net/pdf?id=rJY0-Kcll https://openreview.net/pdf?id=rJY0-Kcll .

FINN C , ABBEEL P , LEVINE S . Model-agnostic meta-learning for fast adaptation of deep networks [EB/OL]. ( 2017-03-09 )[ 2021-09-11 ]. https://arxiv.org/abs/1703.03400 https://arxiv.org/abs/1703.03400 .

LI Z G , ZHOU F W , CHEN F , et al . Meta-SGD: Learning to learn quickly for few-shot learning [EB/OL]. ( 2017-07-31 )[ 2021-09-11 ]. https://arxiv.org/abs/1707.09835 https://arxiv.org/abs/1707.09835 .

LIU T , LUO W , MA L , et al . Coupled network for robust pedestrian detection with gated multi-layer feature extraction and deformable occlusion handling [J]. IEEE Transactions on Image Processing , 2020 , 30 : 754 - 766 .

王程丽 , 赵静 , 杨攀攀 , 等 . 基于迁移学习的小样本SAR船只目标识别 [J]. 移动通信 , 2022 , 46 ( 4 ): 22 - 27 .

WANG C L , ZHAO J , YANG P P , et al . The target recognition of small sample SAR ship based on transfer learning [J]. Mobile Communications , 2022 , 46 ( 4 ): 22 - 27 . (in Chinese)

TAI Y , TAN Y H , XIONG S Z , et al . Few-shot transfer learning for SAR image classification without extra SAR samples [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2022 , 15 : 2240 - 2253 .

高飞 , 赵洁琼 , 林翀 , 等 . 基于距离度量学习的SAR图像识别方法 [J]. 北京理工大学学报 , 2021 , 41 ( 3 ): 334 - 340 .

GAO F , ZHAO J Q , LIN C , et al . A SAR image recognition method based on distance metric learning [J]. Transactions of Beijing Institute of Technology , 2021 , 41 ( 3 ): 334 - 340 . (in Chinese)

WANG L , BAI X R , GONG C , et al . Hybrid inference network for few-shot SAR automatic target recognition [J]. IEEE Transactions on Geoscience and Remote Sensing , 2021 , 59 ( 11 ): 9257 - 9269 .

YANG M J , BAI X R , WANG L , et al . Mixed loss graph attention network for few-shot SAR target classification [J]. IEEE Transactions on Geoscience and Remote Sensing , 2022 , 60 : 1 - 13 .

WANG S Y , WANG Y H , LIU H W , et al . Attribute-guided multi-scale prototypical network for few-shot SAR target classification [J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing , 2021 , 14 : 12224 - 12245 .

FU K , ZHANG T F , ZHANG Y , et al . Few-shot SAR target classification via metalearning [J]. IEEE Transactions on Geoscience and Remote Sensing , 2022 , 60 : 1 - 14 .

HU J , SHEN L , SUN G . Squeeze-and-excitation networks [C]// 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition . Piscataway : IEEE , 2018 : 7132 - 7141 .

WOO S , PARK J , LEE J Y , et al . CBAM: Convolutional block attention module [C]// European Conference on Computer Vision . Cham : Springer , 2018 : 3 - 19 .

Views

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Cross-Scene Point Prediction Crowd Counting Method Based on Meta-Weight-Net

Cross-Modal SAR Target Detection via Progressive Knowledge Transfer

High-Range Resolution Radar Target Recognition Based on Deep Interpretable Probabilistic Auto-Encoder Model

Graph-Based Representation Knowledge Distillation for Image Classification

Related Author

XU Xin

TAN Zhuo-lin

GAO Chen-qiang

XI Yue

ZHAO Guo-wei

JIANG Jia-qing

DONG Gang-gang

LIAO Lei-yao

Related Institution

School of Communications and Information Engineering, Chongqing University of Posts and Telecommunications

Faculty of Science and Technology, University of Macau, Macao

National Key Laboratory of Radar Signal Processing, Xidian University

College of Communication and Information Engineering, Nanjing University of Posts and Telecommunications

Institute of Computing Technology， Chinese Academy of Sciences

⁰