摘要
为了滤除数字图像中同时混有的椒盐噪声和高斯噪声,提出了一种改进的基于灰色关联去混合噪声滤波算法.算法先将滤波窗内噪声点划分两个集合:一个是纯椒盐噪声点和被高斯噪声污染的类椒盐噪声点集合S,一个是被高斯噪声污染的纯高斯噪声点集合P.对滤波窗中心噪声点进行分类平滑处理,如果滤波窗中心点为第一类,用P中像素点的中值作为参考值计算各点的关联系数,作为权值与对应的像素点进行加权计算,结果替换中心噪声点.否则,用P中像素的均值作为参考值计算各点的关联系数,作为权值与对应像素点进行加权计算,结果替换高斯噪声.实验证实,该算法能有效抑制了图像中的混合噪声,提高图像的清晰度,效果优于传统的滤波算法.
In order to filter both salt and pepper noise and gauss noise in digital images,an filtering algorithm for mixed noise based on improved gray correlation is proposed.Firstly,we divide the noise points in the filterwindow into two sets:one is set S:pure salt and pepper noise points and the salt and pepper noise points pollutedby gaussian noise,the other is set P:pure gaussian noise points.The algorithm processes the central noise points separately,and takes the non-salt and pepper noise points in the filtering window as the initial set of pixels P.Ifthe central point of the filtering window is in set S,the correlation coefficients of each point are calculated with the median value of the pixels in S as the reference value,and weighted with the corresponding pixel points.The result replaces the salt and pepper noise point.Otherwise,the correlation coefficient of each point is calculated with the mean value of the pixel in P as the reference value,and weighted with the corresponding pixel as the weight value.The result replaces the gauss noise.Experiments show that the algorithm can effectively suppress the mixed noise in the image and improve the image sharpness,and the effect is better than the traditional filtering algorithms.
作者
沈德海
鄂旭
侯建
阎琦
SHEN Dehai;E Xu;HOU Jian;YAN Qi(College of Information Science and Technology,Bohai University,Jinzhou 121013,China)
出处
《渤海大学学报（自然科学版）》
CAS
2021年第3期259-263,共5页
Journal of Bohai University:Natural Science Edition
基金
国家自然科学基金项目(No:61473045)
辽宁省自然科学基金项目(No:2014020141,No:20170540005)
辽宁省社会科学基金项目(No:L19BGL016)
辽宁省教育科学“十二五”规划2015年度立项课题(No:JG15DB028)
辽宁省教育厅科学研究项目(No:LQ2017003)
关键词
灰色关联
混合噪声
滤波算法
gray relevance
fixed noise
filter algorithm
作者简介
沈德海(1978-),男,副教授,主要从事图像处理方面的研究;通信作者:沈德海,56045499@qq.com.
