车牌字符分割的算法研究与实现  

The Algorithm research and implementation on character segmentation in license plate

作　　者：李志敏[1] 古利超[1] 王克[1] 陈小保[1] 

机构地区：[1]重庆大学光电技术及系统教育部重点实验室

出　　处：《微计算机信息》2012年第9期419-421,共3页Control & Automation

摘　　要：为获得准确的单个车牌字符,提出了一种将连通域法和垂直投影法相结合的车牌字符分割算法。该方法首先根据车牌中每个字母或数字的连通性特点,按照一定的像素准则与相邻像素之间的连通关系扫描目标像素的相邻像素点,找出各自的连通区域;然后对车牌进行垂直投影找出汉字区域,精确地分割出单个字符。实验结果表明,该算法对车牌字符分割的准确性和处理速度都有较大提高,能满足系统实时性和准确性的要求。In order to realize the character segmentation of the target,a character segmentation method of license plates is proposed,which integrates template matching,vertical projection and extracting connected areas method.This method,based on the connectivity features of each letter or number in license plate,scanning the adjacent pixels of target pixels and determining the connectivity of there relationship between them according to several rules,finds their respective connected region.Then it uses the vertical projection method to find the region of the license plate characters.Finally,this method has segmented a great deal characters of license plate.The experiment results indicate that the algorithm given is able to improve the veracity and enhance the speed.It can satisfy real time and veracity requirements of the system.

关 键 词：垂直投影 像素标记 字符分割 字符特征 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]