基于复数信息传递的结构稀疏宽角合成孔径雷达成像算法
第37卷第8期电子与信息学报 Vol.37 No.8 2015年8月 Journal of Electronics & Information Technology Aug. 2015
基于复数信息传递的结构稀疏宽角合成孔径雷达成像算法
蒋成龙*①②③张冰尘①②王正道④洪文①②
①(微波成像技术国家重点实验室北京 100190)
②(中国科学院电子学研究所北京 100190)
③(中国科学院大学北京 100049)
④(爱荷华州立大学美国)
摘要：传统相关处理算法不能完全解决宽角合成孔径雷达(WASAR)成像中目标的散射特性随观测角度变化的问题。稀疏信号处理为该问题提供一种新思路，各向异性问题可以建模成欠定方程组。随角度增大，未知量的规模以观测孔径数目的线性规模增长，导致成功重建难度增大，甚至是重建失败。该文提出一种基于信息传递原理的宽角合成孔径雷达成像方法。根据宽角合成孔径雷达的观测几何及目标散射特性在不同角度之间存在的相关性，建立基于结构稀疏的成像模型；然后利用信息传递原理，提出基于结构稀疏复数信息传递(GCAMP)的成像算法求解该成像模型。仿真结果验证了该方法的有效性。
关键词：合成孔径雷达；宽角合成孔径雷达成像；各向异性散射特性；结构稀疏；信息传递
中图分类号：TN957.52 文献标识码：A 文章编号：1009-5896(2015)08-1793-08 DOI: 10.11999/JEIT141300
Group-sparse Complex Approximate Message Passing Algorithm for Wide Angle Synthetic Aperture Radar Imaging
Jiang Cheng-long①②③ Zhang Bing-chen①② Wang Zheng-dao④ Hong Wen①②
①(Science and Technology on Microwave Imaging Laboratory, Beijing 100190, China)
②(Institute of Electronics, Chinese Academy of Sciences, Beijing 100190, China)
③(University of Chinese Academy of Sciences, Beijing 100049, China)
④(Iowa State University, USA)
Abstract: Conventional matched filtering based algorithms are not sufficiently good at dealing with the anisotropic backscattering behavior of targets in Wide Angle SAR (WASAR) imaging. Sparse signal processing provides a new idea for this problem, the anisotropic problem is modeled as a group of under-determined linear equations. However, the scale of unknowns in the under-determined equations is in linear order of the number of the observation angle.
As the observation angle increases, the anisotropic problem becomes more and more difficult to be solved, even failed for conventional sparse signal processing algorithms. This paper presents a Group- sparse Complex Approximated Message Passing (GCAMP) algorithm for WASAR imaging. Firstly, a group sparse based WASAR imaging model is provided according to the structured property of backscattering coefficients across different observation angles. Secondly, the GCAMP algorithm is derived from the imaging model using message passing theory. Results of simulation demonstrate the effectiveness of the proposed algorithm.
Key words: SAR; Wide Angle SAR (WASAR) imaging; Anisotropic backscattering; Group sparse; Message passing
1引言
宽角合成孔径雷达(Wide Angle SAR, WASAR)[1]指的是合成孔径角度满足使方位向分辨率大于距离向分辨率要求的一类合成孔径雷达观测模式。为实现这类大角度观测，除了需要所观测的
2014-10-11收到，2015-05-11改回，2015-06-09网络优先出版
国家973计划项目(2010CB731905)资助课题
*通信作者：蒋成龙 chenglong.j@https://www.docsj.com/doc/6516908267.html, 回波数据在较长持续时间内保持相干特性之外，同时还要求获得这段时间内雷达平台的精确位置信息，这得益于现代高精度实时的全球定位系统、惯性导航系统。目前主流的WASAR成像算法包括后向投影算法、极坐标格式算法等。由于收集了很大角度范围内观测目标的雷达后向散射信息，所以WASAR可以得到高分辨率的雷达图像，从而提供了常规SAR所不具备的目标识别与可视化的潜力。