摘要
提出了一种基于改进的K-means的车牌字符分割方法。该方法首先利用均值跳变法对车牌区域进行精确定位,再利用改进的K-means算法对车牌字符进行聚类,最后根据K-means算法得到的聚类中心对车牌字符进行分割。实验结果表明,该方法能够准确地分割出车牌字符,且具有较强的抗干扰性。
A character segmentation method of license plate is proposed based on improved K-means algorithm. This method firstly uses the mean jump method to locate the license plate precisely. And the character of license plate is clustered by using improved K-means algorithm. Finally, according to the cluster center, which is acquired by K-means algorithm, segment the character of license plate. The experimental results show that this method can segment the character of license plate accurately, and has a strong anti-interference.
出处
《电视技术》
北大核心
2015年第1期136-138,共3页
Video Engineering
基金
国家重点基础研究发展计划项目(2005CB321901)
软件开发环境国家重点开放实验室开放课题(BUAA-SKLSDE-09KF-03)
