基于神经网络的导体柱单站微波成像

刘昆; 杨了; 夏晴; 唐涛

基于神经网络的导体柱单站微波成像

1.
成都信息工程学院电子工程学院中国气象局大气探测重点实验室, 四川成都 610225
2. 成都信息工程学院大气科学学院, 四川成都 610225

详细信息

作者简介:
刘昆,(1979-),男,安徽人,成都信息工程学院电子工程学院副教授,博士。主要研究方向为计算电磁学、电磁散射与逆散射、优化算法研究;杨了,(1989-),男,四川人,成都信息工程学院电子工程学院硕士生,主要研究方向为计算电磁学;夏晴,(1979-),女,四川人,成都信息工程学院大气科学学院讲师,硕士,主要研究方向为数值计算。

通信作者:
刘昆 E-mail:liukun@cuit.edu.cn

中图分类号: TN957.5
计量
- 文章访问数: 72
- HTML全文浏览量: 27
- PDF下载量: 9
出版历程
- 收稿日期: 2011-11-23
- 网络出版日期: 2020-12-30
- 发布日期: 2012-10-30

Monostatic microwave imaging method for conductor cylinder based on neural network

1.
Electronic Engineering College, Chengdu University of Information Technology Key Laboratory of Atmospheric Sounding, China Meteorological Administration, Chengdu Sichuan 610225, China
2. College of Atmospheric Sciences, Chengdu University of Information Technology, Chengdu Sichuan 610225, China

摘要

摘要: 提出基于神经网络对导体柱目标实现单站微波成像的方法。通过时域有限差分方法获得横磁波(TM)模式下一系列不同形状导体柱目标的散射场,组成神经网络训练样本。再通过组合运用整数微分进化策略和误差反向传播(BP)算法形成混合自学习策略,克服传统BP神经网络收敛速度慢等不足,提高系统自学习能力。最终基于所得神经网络获得二维导体柱微波成像结果。
- 微波成像 /
- 时域有限差分法(FDTD) /
- 神经网络(NN)
Abstract: A microwave imaging method for monostatic using neural network is proposed in this paper.Firstly,the TM mode scattering data of different shapes of conductor cylinder are obtained by finite-difference time-domain(FDTD)method.And all these data are taken as training samples for neural network.A hybrid self-learning strategy based on integral differential evolution strategy and BP algorithm is presented, which can improve self-learning capability of system and overcome the drawbacks of single BP neural network, such as slow astringency.Finally,the satisfactory results of two-dimensional microwave imaging are reached by using the optimized neural network.
- microwave imaging /
- finite-difference time-domain /
- neural network

HTML全文

参考文献(11)

REKANOS I T, RAISANEN A.Microwave imaging in the time domain of buried multiple scatterers by using an FDTD-based optimization technique[J].IEEE Trans Magn,2003, 39(3):1381-1384.

DONELLI M,MASSA A.Computational approach based on a particle swarm optimizer for microwave imaging of two-dimensional dielectric scatterers[J].IEEE Trans Microw Theory Techn,2005,53(5):1761-1776.

刘昆, 廖成, 卫涛, 等.基于时域有限元近场单站导体柱微波成像[J].电波科学学报,2008, 23(5):969-972. LIU Kun,LIAO Cheng,WEI Tao,et al.Near-field monostatic microwave imaging of conductor cylinder based on time-domain finite element[J].Chinese Journal of Radio Science,2008, 23(5):969-972.(in Chinese)

刘昆, 郭在华.介质圆柱的单站微波成像方法[J].电波科学学报, 2011, 26(3):562-566. LIU Kun,GUO Zaihua.A near-field monostatic microwave imaging method for dielectric cylinder[J].Chinese Journal of Radio Science,2011,26(3):562-566.(in Chinese)

QING Anyong.Electromagnetic inverse scattering of multiple perfectly conducting cylinders by differential evolution strategy with individuals in groups(GDES)[J].IEEE Transactions on Antennas and Propagation, 2004, 52(5):1223-1229.

REKANOS I T,RAISANEN A.Microwave imaging in the time domain of buried multiple scatterers by using an FDTD-based optimization technique[J].IEEE Trans Magn, 2003, 39(3):1381-1384.

CUI T J, CHEW W C, YIN X X, et al.Study of resolution and super resolution in electromagnetic imaging for half-space problems[J].IEEE Trans Antennas Propag, 2004, 52(6):1398-1411.

DONELLI M,FRANCESCHINI G,MARTINI A,et al.An integrated multiscaling strategy based on a particle swarm algorithm for inverse scattering problems[J].IEEE Trans Geosci Remote Sensing, 2006, 44(2):298-312.

卫涛.导体目标的超宽带微波成像研究[D].成都:西南交通大学,2008. WEI Tao.Study for Ultra Wide Band Microwave Imaging of Conductor Objects[D].Chengdu:Southwest Jiaotong University, 2008.(in Chinese)

廖成, 卫涛, 陈伟.整数微分进化策略及其在微波成像中的应用[J].西南交通大学学报, 2007, 42(6):647-652. LIAO Cheng, WEI Tao, CHEN Wei. Integer coded differential evolution strategy and application to microwave imaging[J].Journal of Southwest Jiaotong University, 2007, 42(6):647-652.(in Chinese)

CAORSI S, DONELLI M, MASSA A, et al.Detection of buried objects by an electromagnetic method based on a differential evolution approach[C]//Proceedings of the 21st IEEE Instrumentation and Measurement Technology Conference, 2004:1107-1111.