摘要： 提出一种基于模板匹配的最大类间方差车牌字符分割算法。根据字符串的结构和尺寸特征，设计了车牌字符串模板，该模板在车牌区域滑动匹配进行分类，并结合最大类间方差判决准则，确定最佳匹配位置，分割车牌字符。实验表明，该方法自适应性强，可以获得最优的分割效果。

关键词: 车牌字符分割, 模板匹配, 最大类间方差

Abstract: This paper presents a method based on template match and maximum variance between clusters to segment the characters of license plate. The template of character string on license plate is designed, according to the character string’s structural features. The template slips on the license plate to match the right location, and this matching process regarded as a classifying process combines with the judge rule of maximum variance between clusters to get the optimizational matching location to segment the characters.

Key words: License plate characters’ segmentation, Template match, Maximum variance between clusters