一种基于改进模板匹配的车牌字符识别方法  

A Recognition Method of Car License Plate Characters Based on Modified Templet Matching

作　　者：马俊莉[1] 莫玉龙[1] 王明祥[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《小型微型计算机系统》2003年第9期1670-1672,共3页Journal of Chinese Computer Systems

基　　金：国家自然科学基金资助项目 (69875 0 0 9)

摘　　要：提出了一种改进的模板匹配方法 .该方法是在传统的模板匹配方法的基础上 ,通过对字符特征区域的扩大和加强注意设计了一种改进的模板 ,以达到更有效的匹配结果 .将该方法应用于沪宁高速公路收费口处实拍的车牌图像库中 ,其平均识别率达到 97.1% .In this paper, a modified templet matching method is proposed. Based on traditional templet matching method, the modified templet is desighed to achieve a more effective result by extending and strengthening the feature region of the characters. This method is applied to recognize characters in car license plate database of Hu-ning highway and the recognition rate is 97.1%. Experimental results show that the modified templet matching method has higher recognition rate and robustness.

关 键 词：字符识别 模板匹配 图像处理 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]