保存边界的图象平滑方法  

The Method of Preserving Edge Smoothing

作　　者：刘冬[1] 赵荣椿[1] 

机构地区：[1]西北工业大学计算机科学与工程系,陕西西安710072

出　　处：《西北工业大学学报》2000年第1期102-105,共4页Journal of Northwestern Polytechnical University

摘　　要：提出了一种新的图象平滑方法。该方法利用区域模板和线性模板 ,将图象中的高频噪声和合理的边界结构通过采用不同的模板进行平滑 ,达到了既去除噪声 ,又保存图象边界的目的 ,为图象的预处理提供了一种很好的方法。Existing image smoothing methods can not preserve the edges in an image while removing noise. We present a new method——preserving edge smoothing (PES), to solve this problem. We first design several region masks (Fig.2) and linear masks (Fig.3). By using different region masks, we can determine which region a pixel in the image belongs to, then noise can be removed and the region edges can be preserved to a certain degree by substituting a region′s mean value for its pixel′s value. At the same time, we can use linear masks to distinguish the pixels of edges and noise pixels, thus preserving the edges. The experimental results (Figs.7 and 9) show that this method can not only remove noise in image regions, but also preserve the edges.

关 键 词：区域模板 线性模板 图象平滑 边界保存 图象处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TN919.8[自动化与计算机技术—计算机科学与技术]