图像去椒盐噪声滤波器的研究  

Research on Filter for Removing Image Salt and Pepper Noise

作　　者：李小红[1] 蒋建国[1] 吴从中[1] 詹曙[1] 

机构地区：[1]合肥工业大学计算机与信息学院,安徽合肥230009

出　　处：《工程图学学报》2009年第6期107-114,共8页Journal of Engineering Graphics

基　　金：国家自然科学基金资助项目(60474035);国家教育部博士点基金资助项目(20060359004)

摘　　要：在图像平滑过程中,如何滤除噪声同时保护图像的细节是一个研究热点。本文提出并设计一种将十字窗和叉形窗有机结合的多窗口自适应中值滤波器(MWMF滤波器),可根据窗口内图像自身的形状,自适应地选择十字窗和叉形窗。在比较多窗口自适应中值滤波和形态滤波这两种滤波器的优缺点之后,还提出并设计了非线性组合滤波器。实验表明该滤波器滤除椒盐噪声的效果优于单独使用中值滤波或形态学滤噪,在椒盐噪声概率超过0.2时优势尤为明显。How to filter noise but to reserve image details is still a challenge in the image smoothing process. A multi-window adaptive median filter （MWMF） based on cross-window and furcate-window filter is proposed, in which cross window or furcate window can be selected adaptively based on the image shape in the window. Meanwhile, the non-linear filter base on the multi-window median filter and morphology filter is proposed. Experiment indicates that this filter is better than solely using median filter or mathematics morphology in eliminating salt and pepper noise, especially the noise density is denser than 0.2.

关 键 词：计算机应用 图像处理 多窗口自适应中值滤波器 椒盐噪声 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]