颜色与边缘纹理相结合的车牌定位方法  

License plate location method combining color and edge texture

作　　者：陈宏照 谢正光[1] 卢海伦 CHEN Hongzhao;XIE Zhengguang;LU Hailun(School of Electronics and Information,Nantong University,Nantong 226019,China)

机构地区：[1]南通大学电子信息学院,江苏南通226019

出　　处：《现代电子技术》2018年第21期67-70,75,共5页Modern Electronics Technique

基　　金：国家自然科学基金资助项目(61171077)~~

摘　　要：针对单一方式进行车牌定位所存在的问题,提出一种基于OpenCV颜色与纹理相结合的车牌定位方法。该方法首先利用颜色特征进行图像分割;然后利用数学形态法得出候选区域,再对候选区域分析进行粗定位,并对粗定位的结果通过SVM进行精定位;当候选区域没有定位出车牌时,系统将转而利用边缘纹理特征对图像进行定位得出车牌。实验结果表明,该定位方法具有快速、准确率高的优点。A license plate localization method based on color and texture in OpenCV is proposed to solve the problem existing in license plate location method with sole mode.By means of the method,the color feature is used for image segmentation;the mathematical morphological method is used to get the candidate area,which is analyzed for coarse positioning,and then the fine positioning is carried out in combination with the results of coarse positioning by means of SVM;if the license plate isn′t located in the candidate area,the system will use the edge texture feature for image location to get the license plate.A great number of experimental results show this method has the advantages of rapid speed and high accuracy.

关 键 词：车牌定位 SVM 颜色定位 边缘纹理定位 粗定位 精定位 

分 类 号：TN911.73-34[电子电信—通信与信息系统]