基于非局部均值的最优极化舰船检测算法  

Ship Detection Algorithm of Optimal Polarization Based on Non-Local Means

作　　者：陈建宏[1] 赵拥军[1] 刘伟[1] 秦伟 

机构地区：[1]信息工程大学,河南郑州450001 [2]空军作战指挥综合数据库,北京100843

出　　处：《信息工程大学学报》2014年第6期708-713,共6页Journal of Information Engineering University

基　　金：国家自然科学基金资助项目(61302160;41301481)

摘　　要：极化合成孔径雷达数据蕴含了丰富的极化信息,目前已经被应用到海上舰船检测研究。针对现有算法检测过程中存在虚警、漏警和目标分裂等问题,采用非局部均值结合最优极化检测,提出了一种新的极化SAR舰船检测方法。该方法首先采用非局部均值对极化SAR图像进行相干斑抑制;然后提取最优极化特征图,进而分析其统计分布并应用双参数恒虚警检测算法完成极化SAR舰船检测;最后通过Radarsat-2全极化数据对提出算法进行验证。实验结果表明,文章算法能有效抑制相干斑后检测出海上舰船。Polarimetrie synthetic aperture radar data contains rich polarization information, which has been applied to detect ships at sea. Aimed at problems with existing algorithms in the detection process, such as false alarms, leak alarms, and target division, a new method for detecting ships in polarimetric SAR images is proposed by optimal polarization detection after non-local means. The method firstly makes full use of non-local means for despeckling polarimetric SAR images, and ex- tracts the optimal polarization characteristic figure. Subsequently it analyzes its statistical distribution and completes ship detection with two-parameter CFAR algorithms. The proposed algorithm is vali- dated by Radarsat-2 polarimetric SAR data. The experimental results show that the proposed algo- rithm can effectively detect the ships at sea in polarimetric SAR images after suppressing speckles.

关 键 词：合成孔径雷达图像 舰船检测 非局部均值 最优极化 

分 类 号：TP75[自动化与计算机技术—检测技术与自动化装置]