小波包域值法去除合成孔径雷达图像斑点噪声  

Restrain speckle noise of SAR image basedon wavelet package threshold method

作　　者：刘永昌[1] 张平[1] 严卫东[1] 陈华[1] 

机构地区：[1]西北核技术研究所,陕西西安710024

出　　处：《红外与激光工程》2001年第3期160-162,167,共4页Infrared and Laser Engineering

基　　金：国防预研项目的资助 (56 .5 .1 )

摘　　要：合成孔径雷达 (SAR)图像的斑点噪声使图像的解译能力明显降低 ,影响了SAR图像的应用 ,去除图像斑点噪声是SAR图像处理及应用的前提。去除的方法在图像域有很多 ,在频率域 ,也有基于FFT变换和小波变换的许多方法。文中根据沙漠地区的SAR图像的特点 ,对基于图像域和频率域的几种滤波方法进行了比较 ,提出了一种基于小波包变换的SAR图像斑点噪声去除方法 ,结果表明该方法既去除了斑点噪声 ,又保持了边缘特征信息。Speckle noise on synthetic aperture radar image is inevitable. Understanding the SAR image becomes very hard because of the existence of speckle noise, and thus the application of SAR image is influenced severely. There are many methods to restrain the speckle noise in image domain. In frequency domain, there are methods based on FFT transform and WT transform. In this paper, according to the special characteristics of SAR image of desert area, a method based on wavelet package is proposed and compared with other filter methods. The image processing results show that this method restrains the noise effectively, and the edge information of the image is also retained well.

关 键 词：合成孔径雷达 图像处理 噪声 小波变换 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]