基于双重离散小波变换的遥感图像去噪算法  

Remote sensing image de-noising algorithm based on double discrete wavelet transform

作　　者：张倩[1] 

机构地区：[1]河南经贸职业学院信息管理系,郑州450018

出　　处：《国土资源遥感》2015年第4期14-20,共7页Remote Sensing for Land & Resources

基　　金：河南省高等学校教学工程项目"2013年度河南省高等学校教学团队--计算机应用"(编号:教高[2013]589号)资助

摘　　要：结合离散小波变换(discrete wavelet transform,DWT)与二维多级中值滤波(two-dimensional multi-stage median filtering,TMMF)算法,提出了一种遥感图像自适应去噪(de-noising)算法。该算法首先对噪声图像进行单层DWT,分出低频子图像和高频子图像,由于其中的低频子图像包含了大量背景信息,几乎不受噪声的污染,因而变换后仍保持原样无需再处理;对于高频子图像继续执行单层DWT,获得多方向分布的次低频子图像和次高频子图像;然后对次高频子图像采用改进的TMMF算法进行去噪,对次低频子图像采用改进的小波硬阈值函数模型进行噪声抑制;最后对上述去噪后的子图像进行二次小波重构,得到去噪后的遥感图像。采用3景细节信息丰富的遥感图像进行算法性能测试,并与TMMF算法、改进的TMMF算法以及小波变换硬阈值去噪算法进行比较,结果表明,基于双重离散小波变换算法的去噪性能有所提升。In combination with discrete wavelet transform （ DWT ） and two - dimensional multi - stage median filtering（ TMMF） algorithms, the author proposes a self-adaptive remote sensing image de-noising algorithm in this paper. Firstly, the remote sensing noise image is conducted with the single layer DWT so as to obtain the low-frequency wavelet sub-image and high-frequency wavelet sub-images. As the low-frequency sub-image is not polluted by noise, the low-frequency sub-image should be kept unchanged. Signal layer DWT is conducted again for high-frequency sub-images, and therefore the secondary low-frequency sub-image and secondary high-frequency sub-images are obtained. Then, the secondary low-frequency sub-images are filtered by the improved TMMF algorithm, and the secondary high-frequency sub-images are processed by the improved wavelet hard threshold function model. Finally, the wavelet reconstruction image is acquired. Three remote sensing images with detailed information were adopted to test the performance of the method proposed in this paper, and the results of theoretical analysis and test show that the filtering performance of the algorithm proposed in this paper is superior to TMMF algorithm and its improved algorithm as well as wavelet transform hard threshold de-noising algorithm.

关 键 词：遥感图像 离散小波变换(DWT) 二维多级中值滤波(TMMF) 改进的TMMF算法 改进的硬阈值函数模型 去噪算法 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置] TP391.41[自动化与计算机技术—控制科学与工程]