基于Freeman分解的极化SAR数据相干斑滤波

Speckle filtering of PolSAR data based on Freeman decomposition

本文在Freeman分解理论基础上,提出一种极化SAR数据相干斑滤波算法。该算法基于Freeman散射模型对原始极化SAR数据进行分解,以获得像素的散射类型和总功率值;在滤波窗口内选取与中心像素散射类型相同,总功率值接近的像素构成滤波同质区;并根据同质区的局部统计特性,应用线性最小均方滤波器进行滤波处理。试验表明,该算法在有效去除相干斑的同时,对极化和边缘等细节信息也有较好的保持效果。

Abstract： A speckle fihering algurithm based on Freeman deeomposition was proposed in this paper. Firstly, the Freeman decom- position was applied t~ get scattering category and span value of each pixel ; then, homogeneous area was constructed in fihering window by selecting pixels which have the same scattering category and similar span value as central pixel; finally, the LMS filtering was per- formed based on the loeal statistical properties of homogeneous area. Experinmntal resuhs showed that the proposed algorithm could not only filter speckle effectively hut also preserve polarimetric and spatial information well.