基于神经网络和散射中心模型的目标参数提取

doi:10.12305/j.issn.1001-506X.2023.01.02

摘要/Abstract

摘要：

从雷达回波中获取目标几何参数信息往往存在高计算成本、非线性等困难。该文基于卷积神经网络和前馈神经网络, 提出了一种依据散射中心时频像特征的目标类型自动识别和目标几何参数自动提取方法。由于构建一个神经网络需要大量的训练数据样本, 而扩展目标的散射场计算又非常耗时, 利用基于已知目标已建立的散射中心模型, 快速生成大样本训练数据, 有效解决了训练样本难以获得的问题。以弹头类目标为例给出了数值实验结果, 证实了所提方法的有效性。

关键词: 神经网络, 散射中心, 时频特征, 参数提取

Abstract:

Target geometry extraction from radar echoes are often subject to high computational cost, non-linearity, and other difficulties. In this paper, based on convolutional neural network and back propagation neural network, a method is proposed to automatically identify the target pattern and extract the target geometry parameters from the time-frequency image characteristics of scattering center. Since the construction of a neural network requires a large number of training data samples, and the computation of the scattering field of the extended target is very time-consuming, the scattering center model established based on the known target is used in this paper to quickly generate large sample training data, which effectively solves the problem of obtaining training samples. Taking warhead targets as an example, the neural networks are established, and the effectiveness of the proposed method is verified by numerical experiment results.

Key words: neural network, scattering center, time-frequency characteristics, parameter extraction

中图分类号:

TN95

罗宇航, 陈彦锡, 郭琨毅, 盛新庆, 马静. 基于神经网络和散射中心模型的目标参数提取[J]. 系统工程与电子技术, 2023, 45(1): 9-14.

Yuhang LUO, Yanxi CHEN, Kunyi GUO, Xinqing SHENG, Jing MA. Target parameter extraction based on neural network and scattering center model[J]. Systems Engineering and Electronics, 2023, 45(1): 9-14.

图/表 10

图1

表1

图2

图3

表2

表3

图4

图5

表4

表5

参考文献 30

1	LIT L,DUL.Target discrimination for SAR ATR based on scattering center feature and K-center one-class classification[J].IEEE Sensors Journal,2018,18(6):2453-2461. doi: 10.1109/JSEN.2018.2791947
2	DINGB Y,WENG J.Target reconstruction based on 3-D scattering center model for robust SAR ATR[J].IEEE Trans.on Geoscience and Remote Sensing,2018,56(7):3772-3785. doi: 10.1109/TGRS.2018.2810181
3	GUOK Y,QUQ Y.Geometry reconstruction based on attri-butes of scattering centers by using time-frequency representations[J].IEEE Trans.on Antennas and Propagation,2016,64(2):708-720. doi: 10.1109/TAP.2015.2511779
4	LIUJ,HES Y,ZHANGL,et al.An automatic and forward method to establish 3-D parametric scattering center models of complex targets for target recognition[J].IEEE Trans.on Geoscience and Remote Sensing,2020,58(12):8701-8716. doi: 10.1109/TGRS.2020.2989856
5	BHALLAR,MOOREJ.A global scattering center representation of complex targets using the shooting and bouncing ray technique[J].IEEE Trans.on Antennas and Propagation,1997,45(12):1850-1856. doi: 10.1109/8.650204
6	QUQ Y,GUOK Y,SHENGX Q.Scattering centers induced by creeping waves on cone-shaped targets in bistatic mode[J].IEEE Trans.on Antennas and Propagation,2015,63(7):3257-3262. doi: 10.1109/TAP.2015.2424455
7	YANGC X,ZHANGJ,TONGM S.A hybrid quantum-behaved particle swarm optimization algorithm for solving inverse scattering problems[J].IEEE Trans.on Antennas and Propagation,2021,69(9):5861-5869. doi: 10.1109/TAP.2021.3060592
8	WEIY K,LIY C,DINGZ G,et al.SAR parametric super-resolution image reconstruction methods based on ADMM and deep neural network[J].IEEE Trans.on Geoscience and Remote Sensing,2021,59(12):10197-10212. doi: 10.1109/TGRS.2021.3052793
9	PANM A,LIUA L,YUY Z,et al.Radar HRRP target recognition model based on a stacked CNN-Bi-RNN with attention mechanism[J].IEEE Trans.on Geoscience and Remote Sensing,2021, doi: 10.1109/TGRS.2021.3055061
10	MANJUSHAK,KUMARM A,SOMANK P.Integrating scattering feature maps with convolutional neural networks for Malayalam handwritten character recognition[J].Document Analysis and Recognition,2018,21(3):187-198. doi: 10.1007/s10032-018-0308-z
11	FENGS J,JIK F,ZHANGL 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
12	MALMGREN-HANSEND,KUSKA,DALLJ,et al.Improving SAR automatic target recognition models with transfer learning from simulated data[J].IEEE Geoscience and Remote Sensing Letters,2017,14(9):1484-1488. doi: 10.1109/LGRS.2017.2717486
13	MUFTI M A, HADHRAMI E A, TAHA B, et al. Using transfer learning technique for SAR automatic target recognition[C]//Proc. of the SPIE Future Sensing Technologies, 2019.
14	ZHOUY J,LEIZ,CAOY H,et al.Attitude estimation and geometry reconstruction of satellite targets based on ISAR image sequence interpretation[J].IEEE Trans.on Aerospace and Electronic Systems,2018,55(4):1698-1711.
15	WEN Y, CHEN X, WEI Y, et al. SAR parameter estimation method for rectangle plane based on information geometry[C]// Proc. of the IEEE International Conference on Signal, Information and Data Processing, 2019.
16	YUX,EFEM O,KAYNAKO.A general backpropagation algorithm for feedforward neural networks learning[J].IEEE Trans.on Neural Networks,2002,13(1):251. doi: 10.1109/72.977323
17	JIN H. Application of advanced BP neural network in image recognition[C]//Proc. of the 18th International Symposium on Distributed Computing and Applications for Business Engineering and Science, 2019.
18	ZHANGC C,YUS,LIG J,et al.The recognition method of MQAM signals based on BP neural network and bird swarm algorithm[J].IEEE Access,2021,9,36078-36086. doi: 10.1109/ACCESS.2021.3061585
19	郭琨毅,牛童瑶,屈泉酉,等.散射中心的时频像特征研究[J].电子与信息学报,2016,38(2):478-485.
	GUOK Y,NIUT Y,QUQ Y,et al.Study on time-frequency image characteristics of scattering center[J].Journal of Electronics and Information,2016,38(2):478-485.
20	艾俊强,陈如山,陈晓盼,等.电磁计算十大问题[J].电波科学学报,2020,35(1):3-12.
	AIJ Q,CHENR S,CHENX P,et al.Ten problems in electromagnetic calculation[J].Journal of Radio Wave Science,2020,35(1):3-12.
21	殷红成,郭琨毅.目标电磁散射特性研究的若干热点和难点问题[J].电波科学学报,2020,35(1):128-134.
	YINH C,GUOK Y.Some hot and difficult problems in the study of electromagnetic scattering characteristics of targets[J].Journal of Radio Wave Science,2020,35(1):128-134.
22	AUGERF,FLANDRINP.Improving the readability of time-frequency and time-scale representations by the reassignment method[J].IEEE Trans.on Signal Processing,1995,43(5):1068-1089. doi: 10.1109/78.382394
23	XINR Y,JIANGZ,SHAOY T.Complex network classification with convolutional neural network[J].Tsinghua Science and Technology,2020,25(4):447-457. doi: 10.26599/TST.2019.9010055
24	GUO L S, LI M K, XU S H, et al. Study on a recurrent con-volutional neural network based FDTD method[C]//Proc. of the International Applied Computational Electromagnetics Society Symposium, 2019.
25	LAMPRECHT C, BEKHRAD P, IVANOV H, et al. Modelling the refractive index structure parameter: a ResNet approach[C]//Proc. of the International Conference on Broadband Communications for Next Generation Networks and Multimedia Applications, 2020.
26	HAQUE M F, LIM H Y, KANG D S. Object detection based on VGG with ResNet network[C]//Proc. of the International Conference on Electronics, Information, and Communication, 2019.
27	LIU Z M, LI C Q, FENG K. Application research of BP neural network optimized by genetic algorithm and particle swarm optimization algorithm in MBR simulation[C]//Proc. of the 2nd International Conference on Artificial Intelligence and Big Data, 2019.
28	DENGH,ZHANGW X,LIANGZ F.Application of BP neural network and convolutional neural network (CNN) in bearing fault diagnosis[J].IOP Conference Series: Materials Science and Engineering,2021,1043(4):042026. doi: 10.1088/1757-899X/1043/4/042026
29	GUAN C L, YANG Y. Research of computer network security evaluation based on back propagation neural network[C]//Proc. of the IEEE International Conference on Power, Intelligent Computing and Systems, 2019.
30	SHI X X, MA S X, YU C L, et al. Research on prediction model of explosive explosion velocity based on improved BP neural network[C]//Proc. of the IEEE 4th Information Technology, Networking, Electronic and Automation Control Conference, 2020.

散射中心	类型	成因
LSC1	局部型	尖顶曲面反射
LSC4	局部型	边缘绕射
LSC5	局部型	边缘绕射
DSC1	分布型	单曲面反射
DSC2	分布型
DSC3	分布型
DSC4	分布型	平面反射

各层名称	配置	输出大小
Conv1	7×7, 64, stride2	112×112
Conv2_x	$ 3 \times 3 \text{ maxpool, stride 2}\\\left[\begin{array}{ll}3 \times 3 & 64 \\3 \times 3 & 64\end{array}\right] \times 2$	56×56
Conv3_x	$\left[\begin{array}{ll}3 \times 3 & 128 \\3 \times 3 & 128\end{array}\right] \times 2 $	28×28
Conv4_x	$\left[\begin{array}{ll}3 \times 3 & 256 \\3 \times 3 & 256\end{array}\right] \times 2 $	14×14
Conv5_x	$ \left[\begin{array}{ll}3 \times 3 & 512 \\3 \times 3 & 512\end{array}\right] \times 2$	7×7
	average pool, 1000-d fc, softmax	1×1

雷达频率/GHz	信噪比/dB	识别率/%
1	5	94.8
	10	96.2
	15	97.2
2	5	94.5
	10	96.5
	15	97.0
3	5	94.4
	10	96.6
	15	97.2

几何结构	几何参数	相对误差/%
锥体	h₁	7.59
柱体	b₂、h₂	(11.59, 6.63)
锥台	b₁、h₃	(9.76, 12.91)

样本标号	实际参数/m					预测参数/m
样本标号	h₁	h₂	h₃	b₁	b₂	h₁	h₂	h₃	b₁	b₂
1	1.920	2.140	0.171	0.596	0.440	2.126	2.148	0.230	0.537	0.527
2	2.158	1.826	0.189	0.319	0.501	2.118	2.107	0.177	0.315	0.580
3	1.995	1.585	0.100	0.442	0.317	1.757	1.590	0.101	0.509	0.243
4	1.724	1.716	0.189	0.542	0.480	1.926	1.765	0.149	0.539	0.506
5	1.959	2.139	0.219	0.645	0.683	2.191	2.072	0.214	0.557	0.577
6	2.417	2.387	0.196	0.686	0.584	2.201	2.026	0.232	0.636	0.644
7	1.962	1.887	0.184	0.425	0.559	2.035	1.947	0.207	0.499	0.494
8	2.285	1.798	0.143	0.562	0.627	2.209	1.961	0.161	0.567	0.564
9	1.557	2.038	0.257	0.628	0.552	1.653	1.876	0.256	0.557	0.549
10	2.322	2.202	0.156	0.354	0.674	2.296	2.397	0.187	0.425	0.699