基于生成树的图像完全细化算法  

Image fully thinning algorithm based on spanning tree

作　　者：李甦[1] 谭永龙[1] 

机构地区：[1]云南大学信息学院,云南昆明650091

出　　处：《计算机工程与设计》2006年第21期4006-4007,4070,共3页Computer Engineering and Design

基　　金：云南省自然科学基金项目(2005F0012M)

摘　　要：图像细化是图像处理的重要环节,已有的图像细化算法较多,但都存在一些缺陷,限制了算法的使用范围。提出一种基于生成树的图像细化算法,对原图像运用形态学细化算法预处理,对中间结果的连通分支分别建立生成树,利用树的结构特征对图像中各连通分支逐个细化。对指纹图像的实验结果表明,该算法能使图像得到完全细化并能有效的去除毛刺,减小噪声干扰,还能避免交叉点处连通度冗余现象,有较强的适应性。Image thinning is an important segment in the image processing. There are many image thinning algorithms existing, but nearly all of them have some limitations that limit its usage area. Image＇s spanning tree thinning algorithm make the image completely thinning which use the characters of the tree, so it can extract the feature of the image accurately. Firstly, the morphological thinning algorithm pretreatment the image is used, secondly, spanning tree of the image is built, and the characters of the tree to thin the image are used. The test result of the fingerprint image enunciate that image＇s spanning tree thinning algorithm thin the image completely, wipe offthe tiny burr, and avoid the redundancy of connectivity in the crossing. This algorithm has a perfect performance and is suitable to a wide range of image.

关 键 词：细化算法 图像生成树 图像处理 图像细化 骨架化 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]