车牌字符识别技术的研究与实现  

Research and Realization of Vehicle License Plate Character Identification Technology

作　　者：孙炎增[1] 张前进[1] 

机构地区：[1]河南科技大学电子信息工程学院,河南洛阳4710031

出　　处：《微电子学与计算机》2008年第6期101-104,共4页Microelectronics & Computer

基　　金：国家自然科学基金项目(604740356)

摘　　要：设计了一种机动车牌照的自动识别解决方案.对车牌图像的预处理中,提出了一种新的基于非线性均值边缘检测算法,通过大量实验,该检测算法处理后的图像能更好实现车牌的定位与字符分割.在识别系统中,为了更准确地识别字符,使用两种特征提取法提取字符特征,设计了两级分类器对字符进行逐个识别.实验结果表明,这种识别方法具有良好的识别效果,具有较大的实用价值.A method of vehicle license plate recognition is proposed. In the license plate images preprocessing, a new edge detection algorithm based on the non-linear average value is proposed. Many experiments results show that the license plate images detected by above algorithm can be located and segmented easily. And in the recognition system, in order to identify character exactly, two algorithms of features extraction are used to extract the character features. And two poles identification algorithms are designed to identify the character one by one. Experimental results demonstrate the algorithm is feasible and applicable.

关 键 词：边缘检测 车牌定位 字符分割 字符识别 特征提取 

分 类 号：TP391.43[自动化与计算机技术—计算机应用技术]