基于C均值分类的极化SAR图像白化斑点滤波方法

极化合成孔径雷达SAR图像相干斑抑制基本原理是依据各通道间的极化信息差异。提出在各极化通道上分别进行非监督C均值分类,在标号图像中对各区域的水平通道平均散射强度、交叉极化通道与同通道强度比以及同极化通道间的相关系数等参数进行估计,将估计值应用到极化白化滤波器(PWF)中,从而得到改进的PWF滤波输出。经真实极化SAR图像检验,表明该方法能够提高传统PWF滤波器在均匀区域对相干斑的抑制能力。

Polarimetric whitening speckle filter based on C-mean value classification for SAR images

The principle of speckle reduction in polarimetric synthetic aperture radar(SAR) image is using polarimetric diversity among the channels.An improved polarimetric whitening filter(PWF) is proposed.Firstly each channel raw image is classified without surveillance by a C-mean value algorithm,and the label images are merged.Then scatter energy in HH channel,the polarimetric energy ratio between the HV and HH channel and the correlation coefficients of HH and VV images are evaluated in every segmented area,and an improved filtering performance is obtained by using the PWF with these new parameters.Results by the test of real polarimetric SAR images show that in homogeneous areas the proposed method could suppress coherent speckle better than original filters.