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

doi:10.12305/j.issn.1001-506X.2022.01.15

摘要/Abstract

摘要：

针对雷达目标全极化高分辨距离像(high resolution range profile, HRRP)提取可分性特征时, 利用全部距离单元作为度量尺度无法保留各距离单元具体特征的问题, 在综合利用4个极化通道的舰船目标HRRP信息时选择单个距离单元作为度量尺度。在此基础上, 提出基于Pauli分解, H、α、A、α₁分解和结构相似性参数的特征提取方法对目标极化散射矩阵进行特征提取, 并将提取得到的特征与基于卷积神经网络(convolutional neural network, CNN)的舰船目标HRRP识别方法结合, 利用改进残差结构CNN从极化特征中进一步提取深层可分性特征进行目标识别。实验结果表明, 所提方法能够保留目标全极化HRRP更多特征, 提高目标识别的准确率。

关键词: 卷积神经网络, 全极化高分辨距离像, 可分性特征, Pauli分解

Abstract:

Aiming at the problem that when extracting separability features from radar target full polarization high resolution range profile (HRRP), using all range units as the measurement scale cannot retain the specific characteristics of each range unit, a single range unit is selected as the measurement scale when comprehensively using the ship target HRRP information of four polarization channels. On this basis, a feature extraction method based on Pauli decomposition, H、α、A、α₁ decomposition and structural similarity parameters is proposed to extract the features of the target polarization scattering matrix, and the extracted features are combined with the ship target HRRP recognition method based on convolutional neural network (CNN). The improved residual structure CNN is used to further extract deep separability features from polarization features for target recognition. Experimental results show that the proposed method can retain more features of target full polarization HRRP and improve the accuracy of target recognition.

Key words: convolutional neural network (CNN), full-polarization high resolution range profile (HRRP), separable characteristics, Pauli decomposition

中图分类号:

TN957.51

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.

Bo DAN, Zhequan FU, Shan GAO, Tao JIAN. Full-polarization high resolution range profile recognition technology for sea surface target based on convolutional neural network[J]. Systems Engineering and Electronics, 2022, 44(1): 108-116.

图/表 14

图1

图2

图3

表1

图4

图5

图6

图7

表2

图8

图9

图10

图11

表3

参考文献 25

1	范学满, 胡生亮, 贺静波. 对海雷达目标识别中全极化HRRP的特征提取与选择[J]. 电子与信息学报, 2016, 38 (12): 3261- 3268.
	FAN X M , HU S L , HE J B . Feature extraction and selection of full polarization HRRP in target recognition process of maritime surveillance radar[J]. Jounal of Electronics & Information Technology, 2016, 38 (12): 3261- 3268.
2	张玉玺, 王晓丹, 姚旭, 等. 基于复数全极化HRRP的雷达目标识别[J]. 系统工程与电子技术, 2014, 36 (2): 260- 265.
	ZHANG Y X , WANG X D , YAO X , et al. Radar target recognition based on complex fully polarimetric HRRP[J]. Systems Engineering and Electronics, 2014, 36 (2): 260- 265.
3	宋新景. 基于极化特征的雷达目标识别技术[J]. 雷达科学与技术, 2014, 14 (1): 39- 53. doi: 10.3969/j.issn.1672-2337.2014.01.007
	SONG X J . Radar target recognition based on polarization feature[J]. Radar Science and Technology, 2014, 14 (1): 39- 53. doi: 10.3969/j.issn.1672-2337.2014.01.007
4	赵春雷, 王亚梁, 阳云龙, 等. 雷达极化信息获取及极化信号处理技术研究综述[J]. 雷达学报, 2016, 5 (6): 620- 638.
	ZHAO C L , WANG Y L , YANG Y L , et al. Review of radar selection information acquisition and polarimetric signal processing techniques[J]. Journal of Radar, 2016, 5 (6): 620- 638.
5	陈曦, 钟雪莲. 多极化SAR图像的智能识别应用研究[J]. 空军预警学院学报, 2019, 33 (1): 20- 24. doi: 10.3969/j.issn.2095-5839.2019.01.005
	CHEN X , ZHONG X L . Research on intelligent recognition of multipolarization SAR image[J]. Journal of Air Force Early Warning Academy, 2019, 33 (1): 20- 24. doi: 10.3969/j.issn.2095-5839.2019.01.005
6	张胜峰, 朱新国, 马超. 基于Cloude分解的宽带双极化特征提取分析[J]. 电子测量技术, 2018, 41 (24): 99- 102.
	ZHANG S F , ZHU X G , MA C . Wideband dual-polarization feature extraction analysis based on Cloude decompo sition[J]. Jounal of Electronic Measurement Technology, 2018, 41 (24): 99- 102.
7	王宇, 禹卫东, 刘秀清. 基于极化特征参数和极化干涉最优参数的改进四元素分解方法[J]. 电子与信息学报, 2019, 41 (12): 2881- 2888. doi: 10.11999/JEIT190108
	WANG Y , YU W D , LIU X Q . An improved four-component decomposition method based on the characteristic of polarization and the optimal parameters of PolInSAR[J]. Jounal of Electronics & Information Technology, 2019, 41 (12): 2881- 2888. doi: 10.11999/JEIT190108
8	张胜峰, 马超, 朱新国, 等. 弹道目标宽带极化特征提取分析[J]. 现代雷达, 2020, 42 (1): 45- 50.
	ZHANG S F , MA C , ZHU X G , et al. Wideband polarization feature extraction of ballistic targets[J]. Modern Radar, 2020, 42 (1): 45- 50.
9	魏志强, 毕海霞. 基于聚类识别的极化SAR图像分类[J]. 电子与信息学报, 2018, 40 (12): 2795- 2803.
	WEI Z Q , BI H X . PolSAR image classification based on discriminative clustering[J]. Jounal of Electronics & Information Technology, 2018, 40 (12): 2795- 2803.
10	盖晴晴, 韩玉兵, 南华, 等. 基于深度卷积神经网络的极化雷达目标识别[J]. 电波科学学报, 2018, 33 (5): 575- 582.
	GAI Q Q , HAN Y B , NAN H , et al. Polarimetric radar target recognition based on depth convolution neural network[J]. Jounal of Radio Science, 2018, 33 (5): 575- 582.
11	文伟, 曹雪菲, 张学峰, 等. 一种基于多极化散射机理的极化SAR图像舰船目标检测方法[J]. 电子与信息学报, 2017, 39 (1): 103- 109.
	WEN W , CAO X F , ZHANG X F , et al. PolSAR ship detection method based on multiple polarimetric scattering mechanisms[J]. Jounal of Electronics & Information Technology, 2017, 39 (1): 103- 109.
12	HE S H, SUN W F, GUO G R. A method of range-polarization reconstructing of radar target and its application in target recognition[C]//Proc. of the National Aerospace and Electronics Conference, 1994: 93-96.
13	肖怀铁. 宽带极化毫米波雷达目标特征信号测量与识别算法研究[D]. 长沙: 国防科技大学, 2000.
	XIAO H T. Study on measurement and recognition algorithm of wide-band polarized millimeter wave radar target characteristic signal[D]. Changsha: National University of Defense Technology, 2000.
14	CAMERON W , YOUSSEF N , LEUNG L . Simulated polarimetric signatures of primitive geometrical shapes[J]. IEEE Trans.on Geoscience Remote Sensing, 1996, 34 (3): 793- 803. doi: 10.1109/36.499784
15	KROGAGER E . New decomposition of the radar target scattering matrix[J]. Electronics Letters, 1990, 26 (18): 1525- 1527. doi: 10.1049/el:19900979
16	CLOUDE S , POTTIER E . Review of target decomposition theorems in radar polarimetry[J]. IEEE Trans.on Geoscience and Remote Sensing, 1996, 34 (2): 498- 518. doi: 10.1109/36.485127
17	CLOUDE S , POTTIER E . Concept of polarization entropy in optical scattering[J]. Optical Engineering, 1995, 34 (6): 1599- 1610. doi: 10.1117/12.202062
18	BERIZZI F, MARTORELLA M, CAPRIA A, et al. H/α polarimetric features for man-made target classification[C]//Proc. of the IEEE Radar Conference, 2008.
19	张玉玺, 王晓丹, 姚旭, 等. 基于多尺度H/A/α分解的全极化HRRP特征提取[J]. 控制与决策, 2014, 29 (1): 71- 76.
	ZHANG Y X , WANG X D , YAO X , et al. Feature extraction of fully polarimetric HRRP based on multi-scale H/A/α decomposition[J]. Control and Decision, 2014, 29 (1): 71- 76.
20	郭晨, 简涛, 徐从安, 等. 基于深度多尺度一维卷积神经网络的雷达舰船目标识别[J]. 电子与信息学报, 2019, 41 (6): 1302- 1309.
	GUO C , JIAN T , XU C A , et al. Radar HRRP target recognition based on deep multi-scale 1D convolutional neur al network[J]. Journal of Electronics & Information Technology, 2019, 41 (6): 1302- 1309.
21	王容川, 庄志洪, 王宏波, 等. 基于卷积神经网络的雷达目标HRRP分类识别方法[J]. 现代雷达, 2019, 41 (5): 33- 38.
	WANG R C , ZHUANG Z H , WANG H B , et al. HRRP classification and recognition method of radar target based on convolutional neural network[J]. Modern Radar, 2019, 41 (5): 33- 38.
22	刘兴旺. 一种多层预训练卷积神经网络在图像识别中的应用[D]. 武汉: 中南民族大学, 2018.
	LIU X W. The application of a multi-layers pre-training convolutional neural network in image recognition[D]. Wuhan: South-Central University for Nationalities, 2018.
23	付哲泉, 李尚生, 李相平, 等. 基于高效可扩展改进残差结构神经网络的舰船目标识别技术[J]. 电子与信息学报, 2020, 42 (12): 3005- 3012. doi: 10.11999/JEIT190913
	FU Z Q , LI S S , LI X P , et al. Ship target recognition based on highly efficient scalable improved residual structure neural network[J]. Journal of Electronics & Information Technology, 2020, 42 (12): 3005- 3012. doi: 10.11999/JEIT190913
24	YANG J Y , PENG Y N , LIN S M . Similarity between two scattering matrices[J]. Electronics Letters, 2001, 37 (3): 193- 194. doi: 10.1049/el:20010104
25	FU Z Q , LI S S , LI X P , et al. A neural network with convolutional module and residual structure for radar target recognition based on high-resolution range profile[J]. Sensors, 2020, 20 (3): 586. doi: 10.3390/s20030586

参数	取值
中心频率/GHz	10
带宽/MHz	80
采样点数	256
方位角范围/(°)	0~360
方位角间隔/(°)	1
俯仰角/(°)	80
距离单元数	256
分辨单元长度/m	1.875

阶段	输出维度	参数个数
初始卷积层	128×1×9	63+36
卷积模块1	64×1×18	405+180
卷积模块2	32×1×36	1 620+360
卷积模块3	16×1×72	6 480+720
卷积模块4	8×1×144	25 920+1 440
全连接层1	144	0
全连接层2	2	288
输出层	3	6

数据集	参数总数	模型大小/kB	模型复杂度	运算时间/μs	平均识别准确率/%
Data-HH	37 518	772	190 123	155	83.32
Data-HV	37 518	772	190 123	157	80.16
Data-VH	37 518	772	190 123	163	80.10
Data-VV	37 518	772	190 123	161	83.67
Data-HHHVVV	51 414	938	251 017	98	89.86
Data-Pauli	51 414	938	251 017	98	90.08
Data-Full	51 414	938	251 017	106	91.56
Data-Full+Pauli	51 414	938	251 017	112	92.41
Data-Full+H, α, A, α₁	51 414	938	251 017	125	93.05
Data-Full+Similarity	51 414	938	251 017	126	93.74

[1]	宋爽, 张悦, 张琳娜, 岑翼刚, 李浥东. 基于深度学习的轻量化目标检测算法[J]. 系统工程与电子技术, 2022, 44(9): 2716-2725.
[2]	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.
[3]	韦娟, 杨皇卫, 宁方立. 基于NMF与CNN联合优化的声学场景分类[J]. 系统工程与电子技术, 2022, 44(5): 1433-1438.
[4]	陈冬, 句彦伟. 基于语义分割实现的SAR图像舰船目标检测[J]. 系统工程与电子技术, 2022, 44(4): 1195-1201.
[5]	方伟, 王玉, 闫文君, 林冲. 基于神经网络的符号化飞行动作识别[J]. 系统工程与电子技术, 2022, 44(3): 737-745.
[6]	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.
[7]	刘恒燕, 张立民, 闫文君, 钟兆根, 凌青, 梁晓军. 基于WBP-CNN算法的LDPC译码[J]. 系统工程与电子技术, 2022, 44(3): 1030-1035.
[8]	邵凯, 朱苗苗, 王光宇. 基于生成对抗与卷积神经网络的调制识别方法[J]. 系统工程与电子技术, 2022, 44(3): 1036-1043.
[9]	张玺, 金正猛, 姜亚琴. 融合深度图像先验的全变差图像着色算法[J]. 系统工程与电子技术, 2022, 44(2): 385-393.
[10]	金涛, 王晓峰, 田润澜, 张歆东. 基于改进1DCNN+TCN的雷达辐射源快速识别方法[J]. 系统工程与电子技术, 2022, 44(2): 463-469.
[11]	吕勤哲, 全英汇, 沙明辉, 董淑仙, 邢孟道. 基于集成深度学习的有源干扰智能分类[J]. 系统工程与电子技术, 2022, 44(12): 3595-3602.
[12]	高涌荇, 王旭东, 汪玲, 朱岱寅, 郭军, 孟凡旺. 基于RCNN的双极化气象雷达天气信号检测[J]. 系统工程与电子技术, 2022, 44(11): 3380-3387.
[13]	李永刚, 朱卫纲, 黄琼男, 李云涛, 何永华. 复杂背景下SAR图像近岸舰船目标检测[J]. 系统工程与电子技术, 2022, 44(10): 3096-3103.
[14]	刘紫燕, 马珊珊, 梁静, 朱明成, 袁磊. 注意力机制CNN的毫米波大规模MIMO系统信道估计算法[J]. 系统工程与电子技术, 2022, 44(1): 307-312.
[15]	王彩云, 李阳雨, 李晓飞, 王佳宁, 魏文怡. 基于稀疏性深度学习的航拍图像超分辨重构[J]. 系统工程与电子技术, 2021, 43(8): 2045-2050.