车牌识别系统研究与实现  

Study and Realization for License Plate Recognition System

作　　者：李波[1] 曾致远[1] 周建中[1] 罗勤[1] 

机构地区：[1]华中科技大学数字化工程与仿真中心,湖北武汉430074

出　　处：《计算机技术与发展》2006年第6期10-11,14,共3页Computer Technology and Development

基　　金：教育部跨世纪优秀人才培养计划基金资助项目(2003714)

摘　　要：车牌识别(LPR)系统是智能交通系统中的重要组成部分,该系统分为车牌定位、字符切分和字符识别3个模块。文中基于数学形态学方法和边缘特征分析来进行车牌定位,接着进行二值化、引入多指标联合评价函数判断反色等处理,然后基于连通体分析的方法切分字符。实验表明该系统设计方法是可行的。License plate recognition（LPR）system, the important part of the intelligent traffic system, is made up of three parts as license plate orientation, character segmentation and character recognition. In this paper, based on mathematical morphology and edge charaeteristic analysis, license plate orientation is conducted. After binarization and the adoption of multiple indexes evaluation function for color reverse, characters are segmented on the basis of connected component analysis. The experiment indicates that it is feasible to adopt this algorithm in LPR system.

关 键 词：车牌识别 数学形态学 车牌定位 字符切分 图像处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]