一种合成孔径雷达图像舰船假目标检测判别算法  

A False Ship Target Discrimination Algorithm in SAR Image

作　　者：李晨波[1] 刘永信[1] 张晖[1] 王玲琳[1] 

机构地区：[1]内蒙古大学电子信息工程学院,内蒙古呼和浩特010021

出　　处：《中国海洋大学学报（自然科学版）》2017年第2期79-84,共6页Periodical of Ocean University of China

基　　金：国家自然科学基金项目(61362002);海洋公益性科研专项(201505002)资助~~

摘　　要：由于合成孔径雷达(SAR)发射脉冲频率以及成像机理的影响,导致SAR图像在成像时存在方位模糊噪声,严重时会在图像上出现假目标,此时若采用传统的目标检测算法,将会把假目标误判为检测目标,降低目标检测的准确率,影响目标的分类识别。本文分析了假目标产生的原因,从理论上导出了假目标在SAR图像中出现的具体位置。从信号处理的角度,对SAR图像进行方位模糊抑制的仿真。然后采用双参数CFAR检测算法对SAR图像进行目标检测,依据真假目标区域连通面积大小不同,且假目标区域具有不连通、面积小等特点,在目标检测的同时,对真假目标进行判别。最后,采用10m×10m分辨率的SAR图像进行实验验证。实验结果证明该算法可以对由于方位向模糊产生的假目标进行有效检测。Because of the firing pulse frequency, imaging mechanism and other factors cause the azimuth ambigui- ty noise appeared in SAR imaging. There will be false targets appear on this image when the noise is in severe ca- ses. In this case, the traditional target detection algorithm will mistake the false target as detection target, it will reduce the detection accuracy, and it will affect the target classification. The reason of false target appearance is analyzed, specific location decoy appeared in SAR images is deduced. From the point of signal processing view, o- rientation blur suppression simulation of SAR image is implemented. Then double parameter CFAR detection algo- rithm is used for SAR image target detection. Because true and false target connected area size are different, and the false target area is not connected and area size is smaller. At the same time target detection, characteristics of the target area are used to discriminate between true and false targets. Finally, 10 m resolution SAR images are used to verify the target detection. The result of experiment shows that this proposed algorithm can detect the false target caused by the azimuth ambiguity.

关 键 词：合成孔径雷达 假目标 恒虚警率 

分 类 号：TP274[自动化与计算机技术—检测技术与自动化装置]