一种模板匹配和神经网络的车牌字符识别方法  

A Method of Characters in Vehicle Number plate Using Pattern Match and Neural Networks

作　　者：王敏[1] 黄心汉[1] 魏武[1] 李炜[1] 

机构地区：[1]华中科技大学控制科学与工程系

出　　处：《华中科技大学学报（自然科学版）》2001年第3期48-50,共3页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：湖北省重点科学技术发展计划项目!(991PO111);武汉市重点科技攻关计划资助项目!(2 0 0 0 10 10 0 94)

摘　　要：提出了一种基于模板匹配和神经网络的车牌识别方法 .该方法集成了模板匹配识别车牌字符和神经网络识别车牌字符的优势 ,可有效地提高车牌字符的识别率、识别速度和识别系统的泛化能力 .实验结果表明 :大多数情况下 ,该方法的识别率超过 90 % ,识别时间不超过 12 0 0ms ,能有效地识别各种车牌中的字符 。A method of characters in vehicle number plate using pattern match and neural networks is presented. This method integrates the advantages of pattern match and neural networks recognizing characters in vehicle number plate. It can be used to solve at the same time the problem that only pattern match or neural network is difficult to recognize the characters. The recognition rate can be improved with the recognition time reduced and the adaptable ability of recognition method increased. The experimental results show that, for a vehicle number plate, recognition rate of characters is more than 90% and the recognition time of characters is less than 1.2 second by using this method. It is obvious that this method is of more effective recognition ability than other methods.

关 键 词：车牌字符识别 模板匹配 神经网络 集成 识别率 识别速度 识别系统 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]