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

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.