一种改进的小波域均值滤波算法  

An improved wavelet domain mean filtering algorithm

作　　者：胡晋山 付昱凯 康建荣 钟琪 王小兵 HU Jinshan;FU Yukai;KANG Jianrong;ZHONG Qi;WANG Xiaobing(School of Geography,Geomatics and Planning,Jiangsu Normal University,Xuzhou,Jiangsu 221116,China;Sinosteel Maanshan General Institute of Mining Research Co.,Ltd.,Maanshan,Anhui 243000,China)

机构地区：[1]江苏师范大学地理测绘与城乡规划学院,江苏徐州221116 [2]中钢集团马鞍山矿山研究总院股份有限公司,安徽马鞍山243000

出　　处：《测绘科学》2021年第9期55-60,共6页Science of Surveying and Mapping

基　　金：国家自然科学基金项目(41671395,52074133)。

摘　　要：针对传统的小波域均值滤波算法在重建之后会产生马赛克效应的缺陷,提出一种基于形态学开运算改进的小波域均值滤波算法,利用开运算的滤波特性,从几何结构的角度针对性地处理马赛克效应,从而降低马赛克效应对干涉相位图像质量的影响,并将改进的小波域均值滤波算法与传统的小波域均值滤波算法、InSAR干涉相位图去噪中经典的自适应均值滤波、自适应中值滤波、Goldstein滤波以及非局部均值滤波滤波进行实验对比。实验结果表明:改进的小波域均值滤波算法优于其他几种算法。In this paper, the defects of traditional wavelet domain mean filtering algorithm may produce Mosaic effect after reconstruction. An improved wavelet domain mean filtering algorithm based on morphological open operation is proposed. Based on the filtering characteristic of open operation, the Mosaic effect is processed from the Angle of geometric structure to reduce the influence of Mosaic effect on image quality. The improved wavelet domain mean filtering algorithm is compared with the traditional wavelet domain mean filtering algorithm, the classical adaptive mean filtering, adaptive median filtering, Goldstein filtering and Non-local Means for InSAR interferogram denoising. The results show that the improved wavelet domain mean filtering algorithm is superior to other algorithms.

关 键 词：INSAR 滤波算法 数学形态学 小波域均值滤波 

分 类 号：P237[天文地球—摄影测量与遥感]