【摘要】 车牌字符识别是智能车牌识别系统中的核心部分,而车牌字符识别的关键在于提取有效的字符特征.字符的轮廓特征能很好地反应字符图像的局部和全局的结构特征,在分析字符凹凸轮廓结构的基础上,定义了字符凹凸轮廓方向和圈,提出一种基于字符凹凸轮廓特征的字符识别方法.通过对字符凹凸轮廓背景赋值,标记凹凸轮廓方向,提取字符不同凹凸轮廓方向的特征及轮廓特征个数;根据字符轮廓特征及其个数可以对待识别字符粗分类识别;对未识别的字符提取局部特征分类识别.该方法不需要对待识别的字符进行细化,计算量小,识别率高.经实验证实该方法识别效果较好.更多还原

【Abstract】 License plate character recognition is the centre part of Intelligence Vehicle LicensePlate Recognition System,and the key to it is extracting character feature effectively.The outline feature of character can partly or wholly reflect the architectural feature of character,and based on the analysis of concave-convex outline structure of a character,defining the orientation and circle of concave-convex outline,which is a character recognition method based on the concaveconvex outline feature.By assigning to the background of concave-convex outline of a character and marking the orientation of the concave-convex outline,the concave-convex outline orientation feature of a character can be extracted.Then the concave-convex outline orientation feature can be used to classify and recognize the character roughly.To the unrecognized character,local feature should be extracted before classify and recognize.With proposed recognition method,the process of thinning the character to be recognition can be moved,and the calculation is smaller,the recognition rate is higher.更多还原