散射模型的多时相极化SAR斑点噪声滤波算法  

Scattering-model-based speckle filtering method for multi-temporal PolSAR data

作　　者：肖可可[1,2] 汪长城[1] 左廷英[1] 宋才秀[1] 解清华 杨如红 韩潇冰 王增茂 

机构地区：[1]中南大学地球科学与信息物理学院测绘与国土信息工程研究所,长沙410083 [2]湖南省电力勘测设计院,长沙410007

出　　处：《测绘科学》2014年第2期99-103,124,共6页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(40901172);测绘遥感信息工程国家重点实验室开放基金项目(11R03);国家西部测图工程2010年度研究课题;对地观测技术国家测绘局重点实验室开放课题(K201008)

摘　　要：本文提出了一种基于散射模型的多时相极化SAR数据斑点噪声滤波算法,该方法是利用多时相极化SAR数据,在保持像元主散射特性基础上实现斑点噪声抑制。其基本原理是对多时相极化SAR数据的主辅影像进行分类,并通过统计主辅影像中待滤波像元邻域窗口内类别分布情况自适应地选取参与滤波的像素进行滤波。实测数据的实验结果表明,该方法既能够有效地抑制斑点噪声,又能够良好地保持地物散射特性和边缘纹理特征。The speckle noise in SAR image seriously affects the SAR image interpretation and application. To solve the problem, the paper proposed a speckle filtering method named scattering-model-based speckle filtering for multi-temporal polarimetric SAR （PolSAR） data. It can reduce the speckle effectively and preserve the dominant scattering mechanism of targets by using the multi-temporal polarimetric SAR data. The basic idea is to compare the category of central pixel between the master classification map and the slave classification map within a sliding window. The method first applied the unsupervised classification method to classify the master and slave images. Then it selected the pixels involved in the filtering process adaptively based on the statistics of the category distribution of the pixels within the neighborhood window in the master and slave classification maps. The experimental results proved the effectiveness of the algorithm in the speckle reduction.

关 键 词：多时相 极化SAR 斑点噪声 滤波 

分 类 号：P237.3[天文地球—摄影测量与遥感]