一种基于字符轮廓特征的印刷体汉字切分方法  

Printed Chinese Characters Segmentation Based on Characters Contours

作　　者：周泽华[1] 胡学友[1] 

机构地区：[1]合肥学院电子信息与电气工程系,合肥230601

出　　处：《合肥学院学报（自然科学版）》2008年第3期35-38,共4页Journal of Hefei University :Natural Sciences

摘　　要：针对单一汉字切分方法的不足,提出一种基于字符轮廓凹凸特征的印刷体汉字切分方法,首先对扫描后的印刷体汉字文本图像进行二值化处理,根据整篇汉字字符的左右轮廓凹凸特征,把字符图像切成一个个的汉字行,最后根据每一个汉字行的字符串上下轮廓凹凸特征,把每一个汉字行切分为单独的汉字字符.仿真实验表明,提出的方法对于标准的纯印刷体汉字文本图像,可以得到比较理想的切分效果.This paper presents a printed Chinese characters segmentation algorithm which is based on characters contours. First, binarization images come into being when performing binarization process of the scanned picture of printed Chinese characters. Then the character picture is segmented to single character row using the concave and convex of string profiles. With the rule of interval, height and width of standard printed Chinese characters, every character row is segmented to single character which uses the concave and convex of string profiles. Experiments show that this algorithm achieving good separation results for the standard printed Chinese character images are achieved.

关 键 词：凹凸轮廓 行分割 字分割 汉字切分 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]