一种新的车牌图像字符分割与识别算法  

A Novel Algorithm of Character Segmentation in License Plates Recognition

作　　者：陈振学[1] 汪国有[1] 刘成云[2] 

机构地区：[1]华中科技大学图像识别与人工智能研究所图像信息处理与智能控制教育部重点实验室,湖北武汉430074 [2]武汉科技大学信息科学与工程学院,湖北武汉430081

出　　处：《微电子学与计算机》2007年第2期42-44,共3页Microelectronics & Computer

基　　金：总装备部基金项目(51401020201JW0521)

摘　　要：提出了一种新的字符分割算法,该算法利用由中间向两端搜索的方法确定字符的上下边界,且利用一维循环清零法及先验知识垂直分割得到单个的字符。实验结果表明,该算法较好的解决了以上问题,适用于各种质量的车牌图像,分割的正确率较高。The LPR（License Plate Recognition） system consists of three steps： license plate locating, character segmentation and character recognition. Character segmentation is an important step in a LPR system. There are many factors that cause the character segmentation task difficult, such as image noise, plate frame, rivet, space mark, plate rotation and illumination variance. This paper presents a new algorithm for character segmentation. Firstly, chardcters top-bottom edges can be obtained by searching from middle to end. And then, the single characters can be segmented using onedim cycle cleating and prior knowledge. The experiment results show the good performance of the segmentation algorithm.

关 键 词：智能交通系统 字符分割 投影特征 先验知识 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]