基于颜色和边缘信息的车牌定位方法研究  

Research on License Plate Location Method Based on Color and Edge Information

作　　者：郭海富 帅鹏举 王霞 GUO Hai-fu;SHUAI Peng-ju;WANG Xia(Southwest University of Science and Technology,Mianyang,Sichuan 621010,China;Yibin University,Yibin,Sichuan 644000,China)

机构地区：[1]西南科技大学,四川绵阳621010 [2]宜宾学院,四川宜宾644000

出　　处：《科技视界》2018年第5期198-199,205,共3页Science & Technology Vision

摘　　要：为实现车牌的准确定位,本文提出了一种基于车牌颜色和边缘信息结合的车牌定位方法。该方法包括预处理、颜色特征提取、边缘与颜色特征融合,车牌候选区域筛选四个步骤。其二值化使用了一种对光照不敏感的基于颜色特征的方法,解决了车牌区域对光照变化敏感的问题。通过模拟实验,使用不同光照、不同尺寸、不同背景下的130张图片进行车牌定位实验,定位成功率超过95%。In order to achieve the accurate positioning of license plates,this paper proposes a license plate location method based on the combination of license plate color and edge information.The method includes four steps:preprocessing,color feature extraction,edge and color feature fusion,and license plate candidate area screening.The binarization uses a method based on color features that is insensitive to light,and solves the problem of license plate area sensitivity to light changes.Through simulation experiments,130 license plate images with different lighting,different sizes,and different backgrounds were used to perform license plate positioning experiments.The success rate of positioning was over 95%.

关 键 词：车牌定位 颜色信息 数学形态学 边缘检测 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]