摘要： 文章研究了扫描式自动选票机的图象对准。文中提出分级识别方法,采用层次搜索、模板匹配和特征匹配结合方法,进行对准标志第一级识别。采用中心加权套准选举圆圈的第二级识别。实践证明该选票图象自动对准方法是有效的。

关键词: 图象对准, 自动选票机, 对准标志, 模板匹配, 特征匹配, 中心加权套准算法

Abstract: A new hierarchical algorithm for image registration of a new type automatic ballot machine is developed. A special pattern for coordinate positioning is well designed. Both templet matching and feature matching methods are used for first procedure. A central weighted algorithm for voting circle recognition is given for second procedure. The experimental results show that these image registration methods are effective and reliable.

中图分类号: 

• TP391.4