改进自适应中值滤波的图像去噪  

Image noise removal on improvement adaptive medium filter

作　　者：肖蕾[1] 何坤[1] 周激流[1] 吴笛[1] 

机构地区：[1]四川大学计算机学院,成都610065

出　　处：《激光杂志》2009年第2期44-46,共3页Laser Journal

摘　　要：传统自适应中值滤波的最大最小窗口尺寸固定,并且其最大最小窗口相差较大时,运算时间较长,去噪效果并不一定最佳。本文针对传统自适应中值滤波算法的不足,提出了改进自适应中值滤波算法,首先根据椒盐噪声的分布特点,从单幅含椒盐噪声图像中估算出椒盐噪声的浓度,并分析噪声浓度与自适应中值滤波窗口尺寸之间的关系,建立它们之间的函数关系。其次根据噪声浓度确定自适应中值滤波的最大最小窗口尺寸,最后对图像进行自适应中值滤波。实验结果表明本文算法运算时间随着噪声浓度的变化而变化,而且从PSNR角度来看本文去噪效果比传统自适应中值滤波效果较好。Traditional adaptive median filtering performs badly on removing noise and takes a long run time especially when there＇s a big difference between the stable maximum and minimum window size. The algorithm that we have pruposed, first estimates the thickness of asingle image＇s salt - and - pepper noise based on the noise＇s distribution, and it found a functional relation between thickness and window size.After the two window sizes are confirmed by the noise thickness, the ＇algorithm does median filtering on the image. Our experiment shows that the algorithm changes run time according to the change in the noise thickness. With changes in PSNR, we can see that the algorithm performg well to remove the noise.

关 键 词：去噪 自适应中值滤波 最佳窗口尺寸 椒盐噪声 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]