Wishart分布情形下极化SAR图像目标CFAR检测解析方法  

Polarimetric SAR Imagery Target CFAR Detection Analytical Algorithm with Wishart Distribution

作　　者：张嘉峰 朱博 张鹏[1] 王明春 刘涛[1] 

机构地区：[1]海军工程大学电子工程学院,湖北武汉430030 [2]海军装备研究院信息工程技术研究所,北京102200

出　　处：《电子学报》2018年第2期433-439,共7页Acta Electronica Sinica

基　　金：国家自然科学基金(No.61372165)

摘　　要：等效视数是影响极化合成孔径雷达(Pol SAR)图像恒虚警(CFAR)检测性能的重要参数.目前等效视数的估计大都不是整数,导致已有的基于整数等效视数的CFAR检测方法不再适用.为解决此问题,提出了一种新的Pol SAR图像目标CFAR检测解析算法.首先,在Wishart分布假设下,推导出了多视极化白化滤波(MPWF)检测量的概率密度函数;然后对其积分得到了CFAR检测门限关于等效视数的解析表达式;最后通过仿真数据和AIRSAR实测数据比较了新方法与已有的适用于整数等效视数的检测方法和双参数恒虚警(2P-CFAR)检测方法的CFAR检测性能.结果表明新方法中实际虚警概率与给定的恒虚警概率最为接近,更好保证了CFAR检测的恒虚警假设.The equivalent number of looks （ENL） is a key parameter that affects the constant false alarm rate （CFAR） detection performances in polarimetric synthetic aperture radar （PolSAR） imageries. At present,the estimated value of the ENL mostly is not an integer, resulting that existing CFAR detection algorithms are no longer applicable. To solve this problem, a new analytical CFAR algorithm of target detection is proposed in PolSAR imageries. Firstly, the probability density function （PDF） of multi-look polarimetric whitening filter （MPWF） metric is derived based on Wishart distribution assumption; Secondly, the analytical expression of CFAR detection threshold with respect to the ENL is obtained by integra- ting the PDF of MPWF metric; Finally, the CFAR detection performances of the new algorithm are compared with other CFAR detection algorithms via simulation data and AIRSAR measured data. The results show that the real false alarm probability in the new algorithm is the closest to the given constant false alarm probability, which better guarantees the hypothesis of the constant false alarm probability in the CFAR detection.

关 键 词：极化合成孔径雷达 等效视数 WISHART分布 恒虚警检测 解析表达式 

分 类 号：TN95[电子电信—信号与信息处理]