基于压缩感知的SAR图像压缩算法  

SAR image compression algorithm based on compressed sensing

作　　者：许学杰[1,2] 潘志刚[1] 刘畅[1] 

机构地区：[1]中国科学院电子学研究所,北京100190 [2]中国科学院大学,北京100049

出　　处：《中国科学院大学学报（中英文）》2015年第6期790-796,共7页Journal of University of Chinese Academy of Sciences

基　　金：国家自然科学基金(61101201)资助

摘　　要：研究基于压缩感知的合成孔径雷达(SAR)图像压缩算法.根据压缩感知理论,在信号降维方面,提出一种更优化的观测矩阵构造方法;在信号重构方面,提出一种基于微分熵和迭代加权最小二乘的改进重构算法.通过对SAR图像进行压缩和性能比较,得出结论:本文提出的改进算法优于传统的压缩感知算法.In this work,SAR image compression algorithm based on compressed sensing is studied.Based on the theory of compressed sensing,a new approach for measurement matrix construction is proposed in dimension reduction domain while a new algorithm named improved iteratively reweighted least squares based on differential entropy for signal reconstruction is put forward as well.The conclusion can be drawn that the new compression algorithm is prior to traditional compression algorithms after the SAR image compression experiment and the compression performance comparison.

关 键 词：压缩感知 SAR图像压缩 微分熵 迭代加权最小二乘 

分 类 号：TN958[电子电信—信号与信息处理]