改进的基于投影的车牌字符分割算法  

Improved Algorithm for Segmentation of License Plate Characters Based on Projection

作　　者：闫兵[1] 张海朝[1] 

机构地区：[1]河南科技大学,河南洛阳471003

出　　处：《通信技术》2010年第5期184-186,共3页Communications Technology

摘　　要：车牌识别(LPR)是智能交通系统中的一个重要环节,而字符分割是车牌识别中的关键部分。文中提出了一种改进的车牌字符分割算法。首先,该算法结合了数学形态学与二维otsu算法,提高了车牌定位的准确性和抗干扰能力。其次提出了一种新的二值化方法,与传统方法相比,能有效地减少字符粘连,锚点等问题带来的干扰。最后使用投影法分割出字符。实验结果表明:该方法实现简单,分割质量好,效率高,便于下一步的字符识别。License plate recognition is a key process in intelligent transportation system,and characters segmentation of license plate is an important part of vehicle license recognition.This paper presents an improved license plate characters segmentation algorithm,which combines the two-dimensional mathematical morphology and otsu algorithm and improve the accuracy of the positioning and anti-jamming ability.Then a new binarization method is proposed and could effectively reduce the interference caused by adhesions characters.Finally,the characters segmentation is done by projection.Experiments show that the algorithm is simple,the segmentation process is of high quality and efficiency,and good for character recognition.

关 键 词：数学形态学 投影 车牌定位 字符分割 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]