基于纹理特征和颜色匹配的车牌定位方法  

Vehicle License Plate Location Technique Based on Texture Feature and Color Matching

作　　者：黄艳国[1,2] 赵书玲[1] 许伦辉[1] 

机构地区：[1]江西理工大学机电工程学院,江西赣州341000 [2]华南理工大学土木与交通学院,广东广州510640

出　　处：《微电子学与计算机》2011年第9期123-126,共4页Microelectronics & Computer

基　　金：国家自然科学基金项目(60664001);江西省教育厅科技项目(GJJ10481)

摘　　要：在分析车牌定位现有算法的基础上,根据车牌的特点,提出一种新的综合利用车牌纹理特征和边缘颜色对的车牌定位方法.首先根据车牌的纹理特征和结构特点进行粗定位,确定车牌的候选区域,然后对候选车牌区域进行边缘颜色对的检测,根据车牌背景与字符有固定颜色搭配的特点,确定车牌区域.实验结果表明,该算法能有效地对车牌进行定位,提高了车牌定位的可靠性.Based on the analysis of current vehicle license plate location algorithm,according to the characteristic of vehicle plate,a new technique for vehicle license plate location based on texture feature and color is proposed.Firstly the candidate vehicle plate regions are decided through the texture feature and structure of vehicle plate.Secondly the color pattern of the pixels on both sides of the edge point in the candidate vehicle plate regions is investigated,then the vehicle license plate location is extracted correctly from the candidate regions by analyzing the fixed characteristic that the background color matches character color in a license platet.The experiment result shows that the technique can be used to locate vehicle plate region precisely,and the whole performance is improved.

关 键 词：车牌定位 边缘颜色点对 纹理特征 边缘检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]