摘要
字符分割是车牌识别系统的重要步骤。提出一种利用反馈的车牌字符分割算法。首先,对车牌图像进行预处理;其次,根据车牌图像的垂直投影进行字符粗分割;最后,通过两级反馈进行字符精分割。第一级反馈是字符个数与字符宽度反馈,第二级反馈是字符识别反馈。实验结果表明,算法对光照不均、对比度较小、倾斜、污迹、字符粘连和断裂等严重退化的车牌图像具有很好的字符分割性能。
Character segmentation is an important step in license plate recognition system.This paper proposes a segmentation algorithm of license plate character using feedback.Firstly,image pre-processing is applied to the license plate;secondly,vertical projection of the image is used for rough segmentation of the characters on the plate;finally,two-level feedback is employed to conduct fine segmentation of the plate characters.The first level feedback is the feedback based on characters＇ count and the width of single character,the second level feedback is the feedback of character recognition.Experimental results show that the proposed algorithm has good character segmentation performance when dealing with those plates which are uneven-illuminated,contrast–poor,declining,dirt,character-touching and character-broken.
出处
《计算机应用与软件》
CSCD
2011年第4期165-166,283,共3页
Computer Applications and Software
基金
辽宁省博士科研启动基金(20061052)
关键词
车牌字符分割
图像预处理
投影
反馈
离散余弦变换
License plate character segmentation Image pre-processing Projection Feedback Discrete cosine transformation
