复杂背景下单通道SAR运动目标检测方法  

Moving target detection in heterogenous single channel SAR images

作　　者：高飞[1] 周锐[1] 孙进平[1] 

机构地区：[1]北京航空航天大学电子信息工程学院,北京100191

出　　处：《哈尔滨工程大学学报》2011年第3期304-308,共5页Journal of Harbin Engineering University

基　　金：国家自然科学基金资助项目(60702011;61071139)

摘　　要：针对复杂背景和多视情况下合成孔径雷达(SAR)数据相干性较差的问题,提出一种单通道SAR图像中的运动目标检测方法.首先在子视间高重叠度的频域均匀划分基础上,采用三视干涉后对消的杂波抑制;然后基于杂波抑制后图像和参考图像的统计分析作变化检测;最后经恒虚警和形态学处理进一步降低虚警.仿真实验表明,该方法在杂波抑制、降低虚警目标点数等方面具有明显优势.In view of poor coherence in multi-look synthetic aperture radar（SAR） data due to a heterogeneous background,a new technique was presented to highlight moving targets in single channel SAR images.Firstly,clutter suppression through tri-look image cancellation-after-interferometry（CAI） based on frequency even-partition of high level overlapping among sub-looks was performed.Secondly,change detection based on statistical analysis of a clutter-depressed image and reference image was applied.Finally,false alarms were reduced through the standard CFAR（constant false alarm ratio） threshold and proper morphological process.Experimental results were shown to validate the advantages of the proposed method in its clutter suppression and false alarm reduction.

关 键 词：合成孔径雷达 地面运动目标指示 杂波抑制 变化检测 

分 类 号：TN957[电子电信—信号与信息处理]