基于数字图像处理的车牌字符分割方法

An algorithm of characters segmentation based on digital image processing

传统的字符分割方法一般针对单行车牌,本文提出一种针对单行和双行两种结构的车牌字符分割方法。首先进行图像预处理,减少噪声及环境干扰。将车牌图像分为两部分,对前两个字符的部分先垂直投影,再结合先验知识判断进而准确分割;对后五个字符的部分用垂直投影法确定动态阈值并结合连通域分析进行字符分割。实验结果表明,该方法同时适用于单双行结构车牌,对字符粘连断裂的情况也能很好的分割。

An algorithm of characters segmentation is proposed based on digital image processing. In order to accurately carry on the license plate character segmentation,firstly,the positioned licenseplate is preprocessed effectively,and then the license plate image is divided into two parts.One part character uses connected vertical projection combined with prior knowledge of license plate character for the character segmentation.The other part character uses connected domain analysis combined with vertical projection for the character segmentation.The experimental result indicates that this algorithm is the same with single and double row plate.