基于灰度形态学的车牌定位算法  

Algorithm of vehicle license plate location based on grayscale morphology

作　　者：刘军华[1] 雷超阳[1] 

机构地区：[1]长沙通信职业技术学院,湖南长沙410015

出　　处：《交通科学与工程》2012年第4期65-71,共7页Journal of Transport Science and Engineering

基　　金：湖南省科技计划研究项目(2008GK130)

摘　　要：根据车牌图像的统计特征,计算形态学滤波器的结构元素,进行图像背景估计;利用闭-开运算,处理残差图像;采用自动搜索种子区域填充算法,得到各个区域;根据区域的几何特征,判断是否为车牌区域;通过对二值化的车牌采用K-means聚类拟合直线方法进行倾斜校正,得到最终的车牌.实验结果表明:该算法简单、迅速,定位准确率高,为后继字符分割和识别奠定了基础.通过对120幅图像定位实验,总有效定位率为96.7%.用Matlab7.0实验时,每张车牌平均定位时间为2.438s,而用VC++实现时,平均定位时间仅为0.139s.According to the statistical characteristics of the vehicle license plate image, the structure element of a mathematical morphology filter is computed and the image background estimation is performed. The residual image is processed by the close-open operator. All the regions are obtained using region-filling algorithm of automatic search seed. According to the geometry characteristics, the region of vehicle license plate is de- termined. The binarized region of vehicle license plate is carried on tilt correction by the line fitting using K-means clustering, and the final vehicle license plate is obtained. The experimental results show that, this paper algorithm can easily implement and also can quickly and accurately locate the vehicle license plate, which builds a foundation for the successive character segmentation and recognition. The location accuracy on 120 realistic images is up to 96.7%. Under the condition of Matlab7.0, the average processing time of each image is 2. 438 s. But under the condition of VC＋＋6.0, the average processing time of each image is merely 0. 139 s.

关 键 词：车牌号码 灰度形态学 二值化 K—means算法 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]