基于元学习的雷达小样本目标识别方法及改进

doi:10.12305/j.issn.1001-506X.2022.06.09

系统工程与电子技术 ›› 2022, Vol. 44 ›› Issue (6): 1839-1845.doi: 10.12305/j.issn.1001-506X.2022.06.09

基于元学习的雷达小样本目标识别方法及改进

孙晶明^1,^2,*, 虞盛康^1,², 孙俊^1,²

1. 南京电子技术研究所, 江苏南京 210039
2. 中国电子科技集团公司智能感知技术重点实验室, 江苏南京 210039

收稿日期:2021-03-10 出版日期:2022-05-30 发布日期:2022-05-30
通讯作者: 孙晶明
作者简介:孙晶明(1984—), 男, 高级工程师, 博士, 主要研究方向为雷达目标识别、深度学习|虞盛康(1989—), 男, 高级工程师, 博士, 主要研究方向为雷达目标检测与识别、机器学习|孙俊(1974—), 男, 研究员级高级工程师, 博士, 主要研究方向为新体制雷达、雷达信号处理
基金资助:
国家自然科学基金(U19B2031)

Radar small sample target recognition method based on meta learning and its improvement

Jingming SUN^1,^2,*, Shengkang YU^1,², Jun SUN^1,²

1. Nanjing Research Institute of Electronics Technology, Nanjing 210039, China
2. Key Laboratory of IntelliSense Technology, China Electronics Technology Group Corporation, Nanjing 210039, China

Received:2021-03-10 Online:2022-05-30 Published:2022-05-30
Contact: Jingming SUN

摘要/Abstract

摘要：

基于深度学习的雷达目标识别方法近年来获得较大关注, 但实战中存在时效性约束和资源限制, 小样本识别难题大大限制其在实际识别任务中的性能。针对这一问题, 本文基于元学习算法, 通过从多个相关任务中学习到的元知识改善新任务的性能, 引入迁移学习思想, 提出一种改进的小样本学习方法, 并通过详细的性能对比实验分析了该方法的应用边界条件。基于实测高分辨距离像数据的实验结果表明, 元学习方法在历史积累样本所含目标类别较多, 与目标任务相关度较大的极小样本情况下, 性能优势才突出, 所提方法可显著提升其综合识别性能。

关键词: 雷达目标识别, 小样本识别, 元学习, 迁移学习, 高分辨距离像

Abstract:

In recent years, radar target recognition based on deep learning has received great attention. However, there are time constraints and resource constraints in actual combat, and the problem of small sample recognition greatly limits its performance in actual recognition tasks. To solve this problem, based on the meta learning algorithm, this paper improves the performance of new tasks by learning meta knowledge from multiple related tasks, introduces the idea of transfer learning, proposes an improved small sample learning method, and analyzes the application boundary conditions of the method through detailed performance comparison experiments. The experimental results based on the measured high resolution range profile data show that the performance advantage of meta learning method is outstanding only when there are many target categories in the historical accumulated samples and in the case of minimal samples with high correlation with the target task, the proposed method can significantly improve its comprehensive recognition performance.

Key words: radar target recognition, small sample recognition, meta learning, transfer learning, high resolution range profile (HRRP)

中图分类号:

TN957.52

孙晶明, 虞盛康, 孙俊. 基于元学习的雷达小样本目标识别方法及改进[J]. 系统工程与电子技术, 2022, 44(6): 1839-1845.

Jingming SUN, Shengkang YU, Jun SUN. Radar small sample target recognition method based on meta learning and its improvement[J]. Systems Engineering and Electronics, 2022, 44(6): 1839-1845.

图/表 9

图1

图2

图3

图4

表1

表2

图5

图6

图7

参考文献 21

1	LONG T , LIANG Z N , LIU Q H . Advanced technology of high-resolution radar: target detection, tracking, imaging, and recognition[J]. Science China: Information Sciences, 2019, 62, 40301. doi: 10.1007/s11432-018-9811-0
2	LIAO K , SI J X , ZHU F Q , et al. Radar HRRP target recognition based on concatenated deep neural networks[J]. IEEE Access, 2018, 6, 29211- 29218. doi: 10.1109/ACCESS.2018.2842687
3	ZHAO F X , LIU Y X , HUO K , et al. Radar HRRP target recognition based on stacked autoencoder and extreme learning machine[J]. Sensors, 2018, 18 (2): 173- 187. doi: 10.3390/s18010173
4	WAN J W , CHEN B , XU B , et al. Convolutional neural networks for radar HRRP target recognition and rejection[J]. EURASIP Journal on Advances in Signal Processing, 2019,
5	GUO C , HE Y , WANG H P , et al. Radar HRRP target recognition based on deep one-dimensional residual-inception network[J]. IEEE Access, 2019, 7, 9191- 9204. doi: 10.1109/ACCESS.2019.2891594
6	XU B , CHEN B , WAN J W , et al. Target-aware recurrent attentional network for radar HRRP target recognition[J]. Signal Processing, 2019, 155, 268- 280. doi: 10.1016/j.sigpro.2018.09.041
7	DU C , CHEN B , XU B , et al. Factorized discriminative conditional variational auto-encoder for radar HRRP target recognition[J]. Signal Processing, 2019, 158, 176- 189. doi: 10.1016/j.sigpro.2019.01.006
8	袁家雯, 刘文波, 张弓. 基于统计建模的字典学习算法在HRRP的应用[J]. 系统工程与电子技术, 2018, 40 (4): 762- 767.
	YUAN J W , LIU W B , ZHANG G . Application of dictionary learning algorithm in HRRP based on statistical modeling[J]. Systems Engineering and Electronics, 2018, 40 (4): 762- 767.
9	王彩云, 胡允侃, 李晓飞, 等. 基于卷积稀疏编码与多分类器融合的雷达HRRP目标识别方法[J]. 系统工程与电子技术, 2018, 40 (11): 2433- 2437. doi: 10.3969/j.issn.1001-506X.2018.11.07
	WANG C Y , HU Y K , LI X F , et al. Radar HRRP target recognition based on convolutional sparse coding and multi-classifier fusion[J]. Systems Engineering and Electronics, 2018, 40 (11): 2433- 2437. doi: 10.3969/j.issn.1001-506X.2018.11.07
10	潘宗序, 安全智, 张冰尘. 基于深度学习的雷达图像目标识别研究进展[J]. 中国科学: 信息科学, 2019, 49 (12): 1626- 1639.
	PAN Z X , AN Q Z , ZHANG B C . Progress of deep learning-based target recognition in radar images[J]. Scientia Sinica Informationis, 2019, 49 (12): 1626- 1639.
11	韩磊, 姚璐. HRRP自动目标识别方法综述[J]. 北京理工大学学报, 2020, 40 (4): 351- 361.
	HAN L , YAO L . A review of methods for HRRP target automatic recognition[J]. Transactions of Beijing Institute of Technology, 2020, 40 (4): 351- 361.
12	朱克凡, 王杰贵. 小样本条件下SCGAN+CNN低分辨雷达目标一步识别算法[J]. 系统工程与电子技术, 2020, 42 (1): 67- 75.
	ZHU K F , WANG J G . Low-resolution radar target one-step recognition algorithm based on SCGAN+CNN with a limited training data[J]. Systems Engineering and Electronics, 2020, 42 (1): 67- 75.
13	朱克凡, 王杰贵, 刘有军. 小样本条件下基于数据增强和WACGAN的雷达目标识别算法[J]. 电子学报, 2020, 48 (6): 1124- 1131. doi: 10.3969/j.issn.0372-2112.2020.06.012
	ZHU K F , WANG J G , LIU Y J . Radar target recognition algorithm based on data augmentation and WACGAN with a limited training data[J]. Acta Electronica Sinica, 2020, 48 (6): 1124- 1131. doi: 10.3969/j.issn.0372-2112.2020.06.012
14	TAMAAZOUSTI Y , BORGNE H L , HUDELOT C , et al. Learning more universal representations for transfer-learning[J]. IEEE Trans.on Pattern Analysis and Machine Intelligence, 2020, 42 (9): 2212- 2224. doi: 10.1109/TPAMI.2019.2913857
15	HE H , DU L , LIU Y , et al. Similarity preserving multi-task learning for radar target recognition[J]. Information Sciences, 2018, 436/437, 388- 402. doi: 10.1016/j.ins.2018.01.031
16	WANG C W , PEI J F , WANG Z Y , et al. When deep learning meets multi-task learning in SAR ATR: simultaneous target recognition and segmentation[J]. Remote Sensing, 2020, 12 (23): 3863- 3870. doi: 10.3390/rs12233863
17	HUANG Z L , PAN Z X , LEI B . What, where, and how to transfer in SAR target recognition based on deep CNNs[J]. IEEE Trans.on Geoscience and Remote Sensing, 2020, 58 (4): 2324- 2336. doi: 10.1109/TGRS.2019.2947634
18	WEN Y , SHI L C , XU X , et al. HRRP target recognition with deep transfer learning[J]. IEEE Access, 2020, 8, 57859- 57867. doi: 10.1109/ACCESS.2020.2981730
19	KHAN I , ZHANG X C , REHMAN M , et al. A literature survey and empirical study of meta-learning for classifier selection[J]. IEEE Access, 2020, 8, 10262- 10281. doi: 10.1109/ACCESS.2020.2964726
20	YE H J , SHENG X R , ZHAN D C . Few-shot learning with adaptively initialized task optimizer: a practical meta-learning approach[J]. Machine Learning, 2020, 109 (3): 643- 664. doi: 10.1007/s10994-019-05838-7
21	WANG K , ZHANG G , XU Y B , et al. SAR target recognition based on probabilistic meta-learning[J]. IEEE Geoscience and Remote Sensing Letters, 2021, 18 (4): 682- 686. doi: 10.1109/LGRS.2020.2983988

参数名称	参数值
目标型号	P1, P2, P3, P4, P5, Q1, Q2, Q3, Q4
目标信噪比/dB	10~50
目标姿态角/(°)	0~180(样本均匀分布)
样本维度	1×1 024
训练集支撑集样本数	6×16 200
训练集查询集样本数	6×1 800
测试集支撑集样本数	3×N
测试集查询集样本数	3×1 800

实验编号	训练集目标型号	测试集目标型号
1	P4, P5, Q1, Q2, Q3, Q4	P1, P2, P3
2	P1, P2, P3, P4, P5, Q4	Q1, Q2, Q3
3	P3, P4, P5, Q2, Q3, Q4	P1, P2, Q1

[1]	郭军成, 万刚, 胡欣杰, 严发宝, 王帅. 太阳射电频谱爆发识别的元学习方法[J]. 系统工程与电子技术, 2022, 44(8): 2410-2418.
[2]	杨立儒, 刘永祥, 杨威. 基于迁移学习的雷达杂波幅度统计模型选择[J]. 系统工程与电子技术, 2022, 44(8): 2457-2467.
[3]	王彩云, 吴钇达, 王佳宁, 马璐, 赵焕玥. 基于改进的CNN和数据增强的SAR目标识别[J]. 系统工程与电子技术, 2022, 44(8): 2483-2487.
[4]	周晓玲, 张朝霞, 鲁雅, 王倩, 王琨琨. 基于改进R-FCN的SAR图像识别[J]. 系统工程与电子技术, 2022, 44(4): 1202-1209.
[5]	孙晶明, 虞盛康, 孙俊. 基于深度学习的HRRP识别姿态敏感性分析[J]. 系统工程与电子技术, 2022, 44(3): 802-807.
[6]	何立, 沈亮, 李辉, 王壮, 唐文泉. 强化学习中的策略重用: 研究进展[J]. 系统工程与电子技术, 2022, 44(3): 884-899.
[7]	吴涛, 王伦文, 朱敬成. 基于迁移学习和注意力机制的伪装图像分割[J]. 系统工程与电子技术, 2022, 44(2): 376-384.
[8]	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.
[9]	但波, 付哲泉, 高山, 简涛. 基于卷积神经网络的海面目标全极化高分辨距离像识别技术[J]. 系统工程与电子技术, 2022, 44(1): 108-116.
[10]	李睿峰, 许爱强, 孙伟超, 王树友. 基于元学习的航空电子设备特征选择算法推荐方法[J]. 系统工程与电子技术, 2021, 43(7): 2011-2020.
[11]	陈浩, 杨俊安, 刘辉. 基于深度残差适配网络的通信辐射源个体识别[J]. 系统工程与电子技术, 2021, 43(3): 603-609.
[12]	晏媛, 孙俊, 孙晶明, 于俊朋. 雷达小样本目标识别方法及应用分析[J]. 系统工程与电子技术, 2021, 43(3): 684-692.
[13]	胡利平, 董纯柱, 刘锦帆, 殷红成, 王超, 宁超. 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.
[14]	张国令, 吴崇明, 李睿, 来杰, 向前. 基于一维堆叠池化融合卷积自编码器的HRRP目标识别方法[J]. 系统工程与电子技术, 2021, 43(12): 3533-3541.
[15]	张良, 杨威, 李玮杰, 杨小琪, 刘永祥. 基于复数卷积-残差网络的雷达杂波幅度统计模型分类[J]. 系统工程与电子技术, 2021, 43(11): 3086-3097.

基于元学习的雷达小样本目标识别方法及改进

Radar small sample target recognition method based on meta learning and its improvement

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 9

参考文献 21

相关文章 15

编辑推荐

Metrics

本文评价