基于最短欧氏距离匹配的印刷体汉字识别  

The Research on Printed Chinese Characters Recognition Based on Nearest Euclid Distance Template-matching

作　　者：马飞[1] 吕海莲[1] 石果[1] 

机构地区：[1]平顶山学院软件学院,河南平顶山467099

出　　处：《平顶山学院学报》2012年第2期70-73,共4页Journal of Pingdingshan University

基　　金：河南省科技计划重点项目(112102210333);平顶山学院中青年骨干教师专项资助(201138012);平顶山学院高层次人才科研启动经费(2008016)

摘　　要：印刷体汉字识别有着很高的实用价值,主要工作包括图像预处理、字符的定位分割、图像的特征提取和图像识别,并以Visual Studio 2005作为开发环境实现一个完整的基于最短欧氏距离模式匹配的印刷体汉字识别系统.解决了在字符定位分割时出现的"过分离"问题.经测试,识别率可以达到97.6%,基本符合实际应用要求.The recognition of the printed Chinese characters is of practical value.It mainly concerns the image pretreatment,the location and segmentation of the characters,the extraction of the feature and the image recognition,and developes an entity template-matching system based on the nearest Euclid distance,and resolves the over-split problem.Through the tests,the rate of the recognition can reach 97.6%,and it basically meets the needs of practical application.

关 键 词：印刷体汉字 最短欧氏距离 图像识别 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]