噪声伪标签容忍的半监督SAR目标识别

doi:10.12305/j.issn.1001-506X.2025.06.08

摘要/Abstract

摘要：

针对标签训练样本稀缺时半监督合成孔径雷达(synthetic aperture radar, SAR)自动目标识别(automatic target recognition, ATR)中噪声伪标签导致识别精度受限的挑战, 提出一种噪声伪标签容忍的半监督SAR ATR方法。该方法包括两个阶段: 第一阶段通过残差网络(residual network, ResNet)和多分类器融合实现高可靠性伪标签的生成与选择, 从而扩充标签训练数据集; 第二阶段基于WideResNet骨干网络构建具有噪声伪标签容忍特性的鲁棒一致性学习网络, 设计噪声伪标签平滑机制和噪声伪标签容忍的分段损失函数, 实现高精度ATR。在运动和静止目标获取与识别(moving and stationary target acquisition and recognition, MSTAR)SAR数据集上开展实验。实验结果表明, 所提方法在10类目标且每类目标仅有5个标签训练样本的情况下, 能达到93.37%的平均识别准确率, 显著提升了目标识别性能和泛化能力。

关键词: 合成孔径雷达, 自动目标识别, 半监督, 深度学习, 伪标签

Abstract:

To address the challenge of limited recognition accuracy due to noise pseudo-labels in semi-supervised synthetic aperture radar (SAR) automatic target recognition (ATR) with scarce labeled training samples, a noise pseudo-label tolerant semi-supervised SAR ATR method is proposed. The proposed method includes two stages. In the first stage, high-reliability pseudo-labels are generated and selected by combining residual network (ResNet) and multi-classifier fusion, so as to enrich the labeled training dataset. In the second stage, a robust consistency learning network with noise pseudo-label tolerant characteristics is constructed based on WideResNet backbone to implement ATR with high accuracy, in which a noise pseudo-label smoothing mechanism is designed as well as a piecewise noise pseudo-label tolerant loss function. Experiments are conducted on the moving and stationary target acquisition and recognition (MSTAR) SAR dataset. The experimental results demonstrate that the proposed method achieves an average recognition accuracy of 93.37% across 10-class targets with only five labeled training samples for each class, which significantly enhances recognition performance and generalization ability.

Key words: synthetic aperture radar (SAR), automatic target recognition (ATR), semi-supervised, deep learning (DL), pseudo-label

中图分类号:

TP391

张新征, 闫梦可, 朱晓林. 噪声伪标签容忍的半监督SAR目标识别[J]. 系统工程与电子技术, 2025, 47(6): 1796-1805.

Xinzheng ZHANG, Mengke YAN, Xiaolin ZHU. Noise pseudo-label tolerant semi-supervised SAR target recognition[J]. Systems Engineering and Electronics, 2025, 47(6): 1796-1805.

图/表 9

图1

图2

图3

图4

表1

图5

表2

表3

图6

参考文献 48

1	MOREIRA A , PRATS-IRAOLA P , YOUNIS M , et al. A tutorial on synthetic aperture radar[J]. IEEE Geoscience and Remote Sensing Magazine, 2013, 1 (1): 6- 43. doi: 10.1109/MGRS.2013.2248301
2	徐丰, 王海鹏, 金亚秋. 深度学习在SAR目标识别与地物分类中的应用[J]. 雷达学报, 2017, 6 (2): 136- 148.
	XU F , WANG H P , JIN Y Q . Application of deep learning in SAR target recognition and terrain classification[J]. Journal of Radars, 2017, 6 (2): 136- 148.
3	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. doi: 10.1109/JSTARS.2017.2702621
4	XIE Y Z , ZHU B C , YAN W , et al. On finite word length computing error of fixed-point SAR imaging processing[J]. Chinese Journal of Electronics, 2014, 23 (3): 645- 648. doi: 10.23919/CJE.2014.10851225
5	杜兰, 王兆成, 王燕, 等. 复杂场景下单通道SAR目标检测及鉴别研究进展综述[J]. 雷达学报, 2020, 9 (1): 34- 54.
	DU L , WANG Z C , WANG Y , et al. A review of research progress on single-channel SAR target detection and identification in complex scenarios[J]. Journal of Radars, 2020, 9 (1): 34- 54.
6	胡利平, 董纯柱, 刘锦帆, 等. 基于SAR仿真图像的地面车辆非同源目标识别[J]. 系统工程与电子技术, 2021, 43 (12): 3518- 3525. doi: 10.12305/j.issn.1001-506X.2021.12.13
	HU L P , DONG C Z , LIU J F , et al. Heterogeneous target reco- gnition of ground vehicles based on SAR simulated images[J]. Systems Engineering and Electronics, 2021, 43 (12): 3518- 3525. doi: 10.12305/j.issn.1001-506X.2021.12.13
7	罗汝, 赵凌君, 何奇山, 等. SAR图像飞机目标智能检测识别技术研究进展与展望[J]. 雷达学报, 2024, 13 (2): 307- 330.
	LUO R , ZHAO L J , HE Q S , et al. Research progress and prospects of intelligent detection and recognition technology for aircraft targets in SAR images[J]. Journal of Radars, 2024, 13 (2): 307- 330.
8	NOVAK L M , OWIRKA G J , BROWER W S , et al. The automatic target-recognition system in SAIP[J]. Lincoln Laboratory Journal, 1997, 10 (2): 187- 202.
9	张新征, 黄培康. 基于贝叶斯压缩感知的SAR目标识别[J]. 系统工程与电子技术, 2013, 35 (1): 40- 44. doi: 10.3969/j.issn.1001-506X.2013.01.07
	ZHANG X Z , HUANG P K . SAR target recognition based on Bayesian compressed sensing[J]. Systems Engineering and Electronics, 2013, 35 (1): 40- 44. doi: 10.3969/j.issn.1001-506X.2013.01.07
10	ZHAO Q , PRINCIPE J C . Support vector machines for SAR automatic target recognition[J]. IEEE Trans.on Aerospace and Electronic Systems, 2001, 37 (2): 643- 654. doi: 10.1109/7.937475
11	HUANG X , XU Q S , CAO D S , et al. Kernel k-nearest neighbor classifier based on decision tree ensemble for SAR modeling analysis[J]. Analytical Methods, 2014, 6 (17): 6621- 6627. doi: 10.1039/C4AY00836G
12	LOOSVELT L , PETERS J , SKRIVER H , et al. Impact of reducing polarimetric SAR input on the uncertainty of crop classifications based on the random forests algorithm[J]. IEEE Trans.on Geoscience and Remote Sensing, 2012, 50 (10): 4185- 4200. doi: 10.1109/TGRS.2012.2189012
13	HUAN R H, YANG R L. SAR target recognition based on MRF and Gabor wavelet feature extraction[C]//Proc. of the IEEE International Geoscience and Remote Sensing Symposium, 2008.
14	TAO M L , ZHOU F , LIU Y , et al. Tensorial independent component analysis-based feature extraction for polarimetric SAR data classification[J]. IEEE Trans.on Geoscience and Remote Sensing, 2014, 53 (5): 2481- 2495.
15	文贡坚, 朱国强, 殷红成, 等. 基于三维电磁散射参数化模型的SAR目标识别方法[J]. 雷达学报, 2017, 6 (2): 115- 135.
	WEN G J , ZHU G Q , YIN H C , et al. SAR target recognition method based on 3D electromagnetic scattering parametric model[J]. Journal of Radars, 2017, 6 (2): 115- 135.
16	齐会娇, 王英华, 丁军, 等. 基于多信息字典学习及稀疏表示的SAR目标识别[J]. 系统工程与电子技术, 2015, 37 (6): 1280- 1287.
	QI H J , WANG Y H , DING J , et al. SAR target recognition based on multi-information dictionary learning and sparse repre sentation[J]. Systems Engineering and Electronics, 2015, 37 (6): 1280- 1287.
17	ZHAO Z Y , XUE X R , MARIAM I , et al. Integrating target and shadow features for SAR target recognition[J]. Sensors, 2023, 23 (19): 8031. doi: 10.3390/s23198031
18	李毅, 杜兰, 周可儿, 等. 基于属性散射中心卷积核调制的SAR目标识别深层网络[J]. 雷达学报, 2024, 13 (2): 443- 456.
	LI Y , DU L , ZHOU K E , et al. Deep network for SAR target recognition based on attribute scattering center convolutional Kernel modulation[J]. Journal of Radars, 2024, 13 (2): 443- 456.
19	WAGNER S A . SAR ATR by a combination of convolutional neural network and support vector machines[J]. IEEE Trans.on Aerospace and Electronic Systems, 2016, 52 (6): 2861- 2872. doi: 10.1109/TAES.2016.160061
20	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. doi: 10.1016/j.patcog.2016.05.028
21	SHANG R H , WANG J M , JIAO L C , et al. SAR targets classification based on deep memory convolution neural networks and transfer parameters[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2018, 11 (8): 2834- 2846. doi: 10.1109/JSTARS.2018.2836909
22	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. doi: 10.1016/j.isprsjprs.2020.01.016
23	刘旗, 张新禹, 刘永祥. 基于门控多尺度匹配网络的小样本SAR目标识别[J]. 系统工程与电子技术, 2022, 44 (11): 3346- 3356. doi: 10.12305/j.issn.1001-506X.2022.11.08
	LIU Q , ZHANG X Y , LIU Y X . Few-shot SAR target recognition based on gated multi-scale matching network[J]. Systems Engineering and Electronics, 2022, 44 (11): 3346- 3356. doi: 10.12305/j.issn.1001-506X.2022.11.08
24	LI W , GAO Y H , ZHANG M M , et al. Asymmetric feature fusion network for hyperspectral and SAR image classification[J]. IEEE Trans.on Neural Networks and Learning Systems, 2023, 34 (10): 8057- 8070. doi: 10.1109/TNNLS.2022.3149394
25	徐丰, 金亚秋. 微波视觉与SAR图像智能解译[J]. 雷达学报, 2024, 13 (2): 285- 306.
	XU F , JIN Y Q . Microwave vision and intelligent interpretation of SAR images[J]. Journal of Radars, 2024, 13 (2): 285- 306.
26	崔宗勇, 杨致远, 蒋阳, 等. 面向SAR目标识别深度网络可理解的类激活映射方法[J]. 雷达学报, 2024, 13 (2): 428- 442.
	CUI Z Y , YANG Z Y , JIANG Y , et al. Interpretable class activation mapping method for SAR target recognition deep networks[J]. Journal of Radars, 2024, 13 (2): 428- 442.
27	康妙, 计科峰, 冷祥光, 等. 基于栈式自编码器特征融合的SAR图像车辆目标识别[J]. 雷达学报, 2017, 6 (2): 167- 176.
	KANG M , JI K F , LENG X G , et al. SAR image vehicle target recognition based on stacked autoencoder feature fusion[J]. Journal of Radars, 2017, 6 (2): 167- 176.
28	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 Trans.on Geoscience and Remote Sensing, 2020, 59 (3): 2174- 2187.
29	FENG S J , JI K F , ZHANG L B , et al. SAR target classification based on integration of ASC parts model and deep learning algorithm[J]. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14, 10213- 10225. doi: 10.1109/JSTARS.2021.3116979
30	AI J Q , QU Z , ZHAO Z C , et al. A SAR target classification algorithm based on the central coordinate attention module[J]. IEEE Sensors Journal, 2024, 24 (2): 1941- 1952. doi: 10.1109/JSEN.2023.3338218
31	LEE D H. Pseudo-label: the simple and efficient semi-supervised learning method for deep neural networks[C]//Proc. of the Workshop on Challenges in Representation Learning, 2013.
32	SOHN K , BERTHELOT D , CARLINI N , et al. Fixmatch: simplifying semi-supervised learning with consistency and confidence[J]. Advances in Neural Information Processing Systems, 2020, 33, 596- 608.
33	WANG Y D, CHEN H, HENG Q, et al. Freematch: self-adaptive thresholding for semi-supervised learning[C]//Proc. of the International Conference on Learning Representations, 2023.
34	RIZVE M N, DUARTE K, RAWAT Y S, et al. In defense of pseudo-labeling: an uncertainty-aware pseudo-label selection framework for semi-supervised learning[C]//Proc. of the International Conference on Learning Representations, 2021.
35	ZHANG B W , WANG Y D , HOU W X , et al. Flexmatch: boosting semi-supervised learning with curriculum pseudo labeling[J]. Advances in Neural Information Processing Systems, 2021, 34, 18408- 18419.
36	CHEN H, TAO R, FAN Y, et al. Softmatch: addressing the quantity-quality trade-off in semi-supervised learning[C]//Proc. of the International Conference on Learning Representations, 2023.
37	ZHENG C , JIANG X , LIU X Z . Semi-supervised SAR ATR via multi-discriminator generative adversarial network[J]. IEEE Sensors Journal, 2019, 19 (17): 7525- 7533. doi: 10.1109/JSEN.2019.2915379
38	WANG C , SHI J , ZHOU Y Y , et al. Semi-supervised learning-based SAR ATR via self-consistent augmentation[J]. IEEE Trans.on Geoscience and Remote Sensing, 2020, 59 (6): 4862- 4873.
39	WANG C W , PEI J F , YANG J Y , et al. Recognition in label and discrimination in feature: a hierarchically designed lightweight method for limited data in SAR ATR[J]. IEEE Trans.on Geoscience and Remote Sensing, 2022, 60, 5239613.
40	ZHANG X Z , LUO Y Q , HU L P . Semi-supervised SAR ATR via epoch-and uncertainty-aware pseudo-label exploitation[J]. IEEE Trans.on Geoscience and Remote Sensing, 2023, 61, 5209015.
41	ZHANG L, YANG M, FENG X C. Sparse representation or collaborative representation: which helps face recognition?[C]// Proc. of the International Conference on Computer Vision, 2011.
42	WANG B, LI W F, POH N, et al. Kernel collaborative representation-based classifier for face recognition[C]//Proc. of the IEEE International Conference on Acoustics, Speech and Signal Processing, 2013.
43	CUBUK E D, ZOPH B, SHLENS J, et al. Randaugment: practical automated data augmentation with a reduced search space[C]//Proc. of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops, 2020.
44	BERTHELOT D, CARLINI N, GOODFELLOW I, et al. Mixmatch: a holistic approach to semi-supervised learning[C]//Proc. of the Conference on Neural Information Processing Systems, 2019.
45	RODRÍGUEZ P, LARADJI I, DROUIN A, et al. Embedding propagation: smoother manifold for few-shot classification[C]// Proc. of the Computer Vision-ECCV 16th European Conference, 2020: 23-28.
46	ZHOU X, LIU X M, JIANG J J, et al. Asymmetric loss functions for learning with noisy labels[C]//Proc. of the International Conference on Machine Learning, 2021.
47	HUMMEL S R, DARPA Z. The moving and stationary target acquisition and recognition (MSTAR) program[C]//Proc. of the IEEE Workshop on Computer Vision beyond the Visible Spectrum, 1999.
48	LOSHCHILOV I, HUTTER F. SGDR: stochastic gradient descent with warm restarts[C]//Proc. of the International Conference on Learning Representations, 2017.

类别	样本数量(俯仰角为17°)	样本数量(俯仰角为15°)
2S1	299	274
BMP2	233	195
BRDM2	298	274
BTR60	256	195
BTR70	233	196
D7	299	274
T62	299	273
T72	232	196
ZIL131	299	274
ZSU234	299	274
总计	2 425	2 747

算法	每类标签样本数量
算法	5	10	20	40
Fixmatch^[32]	30.65	93.41	96.51	98.19
MGAN^[37]	—	—	85.23	90.82
SCA^[38]	—	95.22	98.23	99.85
HDLM^[39]	72.37	88.16	95.17	97.85
EUAPS^[40]	88.65	98.58	99.84	99.86
本文方法	93.37	99.33	99.86	99.90

伪标签预测	噪声平滑模块	分段损失函数	样本数量
伪标签预测	噪声平滑模块	分段损失函数	5	10
－	－	－	63.26	95.13
√	－	－	91.75	98.56
－	√	－	76.41	98.47
－	－	√	81.03	98.39
√	√	√	93.37	99.33

[1]	黄丰卓, 冯东, 华洋晟, 黄晓涛. 一种改进的全息SAR相位梯度自聚焦算法[J]. 系统工程与电子技术, 2025, 47(6): 1786-1795.
[2]	贾竣皓, 陈学斌, 李璋峰, 叶春茂. 空间群目标ISAR成像与检测方法[J]. 系统工程与电子技术, 2025, 47(6): 1843-1854.
[3]	姜智杰, 宋恒, 胡楠, 段兰茜, 曹平. 隧道环境毫米波雷达目标识别与分类算法[J]. 系统工程与电子技术, 2025, 47(5): 1453-1460.
[4]	刘宁, 李芳芳, 李新武, 洪文. 结合足迹和相位信息的SAR高层建筑三维重建[J]. 系统工程与电子技术, 2025, 47(5): 1469-1486.
[5]	聂千祁, 沙明辉, 朱应申, 王崇宇, 崔念强. 基于盲源分离结合奇异谱分析的雷达多分量信号识别方法[J]. 系统工程与电子技术, 2025, 47(4): 1168-1175.
[6]	袁磊, 杨艳娟, 郭毅, 戴鹏. 相关噪声下基于深度学习的LDPC码码率半盲识别算法[J]. 系统工程与电子技术, 2025, 47(4): 1335-1345.
[7]	贾钰嘉, 张思乾, 唐涛, 匡纲要. 强化散射特征的机载SAR实传图像盲超分辨重建[J]. 系统工程与电子技术, 2025, 47(3): 753-767.
[8]	杨志伟, 杨安东, 梁庚辰, 李相海, 李晓蕊, 刘杰. 多子带融合解动目标径向速度模糊方法[J]. 系统工程与电子技术, 2025, 47(3): 788-796.
[9]	陈世龙, 刘霖, 王晓蓓, 曾翰森, 刘亚波, 刘翔. 脉间捷变频SAR快速补偿频域成像处理算法[J]. 系统工程与电子技术, 2025, 47(3): 797-806.
[10]	沈夏闰, 李若楠, 张昊田. 基于CVAE-LSTM的服务器KPI异常检测[J]. 系统工程与电子技术, 2025, 47(3): 1019-1027.
[11]	贾蕾蕾, 刘利民, 董健. 基于图像结构信息的可见光和SAR图像快速配准[J]. 系统工程与电子技术, 2025, 47(2): 428-441.
[12]	刘晓琳, 郭梦娇, 李卓. Dueling DQN优化下的航班延误自适应图卷积循环网络预测方法[J]. 系统工程与电子技术, 2025, 47(2): 568-579.
[13]	付卫红, 张鑫钰, 刘乃安. 基于多尺度融合神经网络的同频同调制单通道盲源分离算法[J]. 系统工程与电子技术, 2025, 47(2): 641-649.
[14]	蒋李兵, 杨庆伟, 郑舒予, 王壮. 基于拍卖理论的组网雷达多轨道目标ISAR成像资源分配算法[J]. 系统工程与电子技术, 2025, 47(1): 81-93.
[15]	孟洋, 周国如, 李洁, 张冰尘. 基于结构化字典学习的判别稀疏微波成像方法[J]. 系统工程与电子技术, 2025, 47(1): 94-100.