一种高分辨率合成孔径雷达并行成像实现  

Implementation for High Resolution SAR Parallel Imaging

作　　者：刘筱[1] 皮亦鸣[1] 冷传航[2] 

机构地区：[1]电子科技大学电子工程学院,四川成都610054 [2]中国运载火箭技术研究院研发中心,北京100076

出　　处：《信息与电子工程》2008年第1期24-28,共5页information and electronic engineering

基　　金：教育部新世纪优秀人才支持计划资助项目(23901019);四川省青年科技基金会资助项目(06ZQ026-006)

摘　　要：并行处理是快速处理合成孔径雷达数据(SAR),实现实时处理的有效途径。本文采用了一种细粒度的并行算法实现SAR成像,将原始数据划分为子孔径数据,并分配给并行机的各节点,使用针对子孔径的CS处理和方位向子孔径算法,对各节点的数据进行处理。通过在并行机SGIOrigin2000平台上的实验证明,该算法与中粒度并行CS成像算法相比,能有效地减少通信量,有利于成像处理的并行扩展,同时获得更好的分辨率。Parallel processing is an effective way to process Synthetic Aperture Radar （SAR） data quickly and realize real-time imaging. In this paper, a kind of fine-grained parallel algorithm is applied. The raw SAR data is divided to subaperture data which are distributed to the nodes of parallel computer. Chirp Scaling（CS） processing aimed to subaperture data and subaperture processing is used to imaging. Through the experiments on SGI Origin2000, it is proved that, compared with medium-grained CS algorithm, the method presented in this paper reduces the communicational requirement and is good for parallel scalability of imaging as well as achieving better resolution.

关 键 词：合成孔径雷达 并行成像 细粒度 子孔径 高分辨率 

分 类 号：TN958[电子电信—信号与信息处理]