一种基于最优模块匹配的新型加权中值滤波算法  

A New Weighted Median Filtering Algorithm Based on Optimal Block Matching

作　　者：秦晓波[1] 赵国栋[2] 

机构地区：[1]贵阳学院数学与信息科学学院,贵州贵阳550005 [2]上海电机学院数学与物理研究所,上海201306

出　　处：《测控技术》2013年第12期24-26,31,共4页Measurement & Control Technology

基　　金：上海市高校选拔培养优秀青年教师科研专项基金项目(SDJ10015);贵州省科技厅;贵阳市科技局;贵阳学院三方联合基金项目(LKG[2013]49)

摘　　要：为了去除图像中的噪声,提出了一种新型的加权滤波算法,算法思想是基于大范围的最优模块匹配,根据模块之间的相似度来衡量像素点之间的相关性,从而确定对中心点像素贡献的大小,赋予像素点不同大小的权值,来恢复被噪声点污染的信号点,达到去除噪声、保护图像边缘信息的目的。实验表明新算法在滤波性能方面优于现有的一些改进算法,即使在高密度噪声下,也能取得较好的滤波效果。In order to remove the image noise,a new type of weighted filtering algorithm is presented.The idea is proposed based on a wide range of optimal matching module.According to the similarity of module,the correlation between pixels is measured to determine the size of the contribution of the center pixel.With giving different weights and restoring signal point polluted by noise,the purposes of removing the noise and protecting the image edge information are achieved.Experiments show that the new algorithm is superior to some of the existing improved algorithms in filtering performance,even in high-density noise,also to achieve better filtering effect.

关 键 词：模块匹配 加权中值 相关性 去噪 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]