退化字符图象的骨架法形态分析识别方法  

Topographic Approach Recognizer for Skeletonization of Regraded Character Images Based on Fuzzy Set

作　　者：卢达[1] 浦炜[1] 谢铭培[2] 

机构地区：[1]常熟高等专科学校,常熟215500 [2]复旦大学计算机科学系,上海200433

出　　处：《计算机工程》1999年第9期32-34,共3页Computer Engineering

基　　金：1997年度江苏省教委留学回国人员科研基金

摘　　要：基于模糊集合的退化字符图象的骨架法结构形态分析识别方法，可直接应用于退化字符的模糊图象或灰度罗象，因此对噪声不敏感，而对笔划缺损字符可正确地提取整个骨架；处理连体字符时能根据骨架的形态结构特征较容易地确定切分位置，经修剪算法对各切分点的处理，被分离的字符骨架可直接通过分类器而不必再作进一步处理。A topographic approach recognizer is proposed for skeletonization of regraded character images based on fuzzy set. Themeithod worki directly on fuzzy or gray-scale images of regraded character, so that it is less sensitive to noise. It can correctly extract thewholc skeleton although a character is broken into pieces. For some characters merged together, the breaking poritions can be easily locatedby topographic approaching for skeletons. A splithng algorithm is then applied on each breaking position. The separated skeletons can bedbecdy opssed through a classitier without any further process.

关 键 词：退化字符图象 骨架法 图象识别 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]