复杂背景下的车牌自动识别系统  

Vehicle License Plate Recognition System under Complex Background

作　　者：陈梅[1] 

机构地区：[1]曲阜师范大学电气信息与自动化学院

出　　处：《电子技术（上海）》2009年第8期35-36,共2页Electronic Technology

摘　　要：提出一种改进的模糊C均值聚类算法用来对车牌图像进行分割,算法中通过图像的灰度直方图来初始化聚类中心与聚类数目,并对聚类中的隶属度做了相应的修正。车牌的定位是根据水平灰度值的变化规律来实现的;字符的分割是根据字符区域中字符像素个数的垂直投影实现的。实验结果表明该算法能够获得较理想的车牌自动识别效果。In this paper, an improved FCM clustering algorithm is proposed to segment the vehicle license plate image. In the algorithm, the cluster center and cluster number is initialized by the histogram of the image gray-scale, and the membership in clustering is amended. The location of the license plate is realized according to the change of the horizontal gray value. And the character segmentation is realized according to the vertical projection of the number of pixels. Experiment results show that the algorithm is effective to the license plate recognizing system.

关 键 词：模糊C均值聚类算法 图像分割 车牌定位 字符分割 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构] U491[自动化与计算机技术—计算机科学与技术]