首页 | 本学科首页   官方微博 | 高级检索  
     

基于HSI空间的颜色算法提牌照识别的性能
引用本文:傅一平,李志能,袁丁.基于HSI空间的颜色算法提牌照识别的性能[J].计算机工程与设计,2004,25(5):703-707.
作者姓名:傅一平  李志能  袁丁
作者单位:浙江大学,信电系,浙江,杭州,310027
摘    要:为了研制高性能的车辆牌照识别系统,在原有基于灰度图牌照分析的基础上,考虑利用牌照的颜色信息来提高牌照识别的性能,提出了HSI空间上确定彩色牌照颜色的方法,进而用改进的基于K-means的HIS空间颜色聚类算法得到背景色范围,然后用数字形态学方法对聚类后的牌照背景进行一系列的形态运算,去除聚类后得到的前景字符对背景的干扰,最后,用边界颜色拟和直线确定精确的牌照边界,并实现对倾斜牌照的矫正。对大量的彩色车牌图片进行实验结果表明,利用HSI空间的牌照颜色信息可以提高整体牌照识别系统的性能。

关 键 词:HSI空间  牌照识别  颜色聚类算法  数字形态学  光学字符识别
文章编号:1000-7024(2004)05-0703-05

Improving license plate recognition system by color algorithms based on HIS space
FU Yi-ping,LI Zhi-neng,YUAN Ding.Improving license plate recognition system by color algorithms based on HIS space[J].Computer Engineering and Design,2004,25(5):703-707.
Authors:FU Yi-ping  LI Zhi-neng  YUAN Ding
Affiliation:FU Yi-ping,LI Zhi-neng,YUAN Ding Department of Information and Electronic Engineering,Zhejiang University Hangzhou 310027,China
Abstract:In order to improve the performance of license plate recognition system,the color of license plate was considered fol- lowing the analysis of the gray.First,a new method for determining the color of license plate was presented based on HIS color space.And then an improved k-means color clustering algorithm on HIS space was provided to position the background of license plate.And then by further morphological operations,the forground characters which blured the background was removed.Finally, boundary lines of license plate based on the background were determined,which rectified the lean license plate.Experiment result shows that the color algorithms of license plate based on HIS space can improve the whole performance of license plate recognition system greatly.
Keywords:color clustering  mathematical morphology  recognition of license plate
本文献已被 CNKI 维普 万方数据 等数据库收录!
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司  京ICP备09084417号