基于模板匹配的人民币纸币序列号识别系统研究  

Research on RMB Currency Number Identification System Based on Template Matching

作　　者：张冬娇 甘早斌[2] 

机构地区：[1]武汉市石牌岭职业高级中学,武汉430070 [2]华中科技大学计算机科学与技术学院,武汉430074

出　　处：《计算机与数字工程》2012年第1期104-108,共5页Computer & Digital Engineering

基　　金：国家自然科学基金(编号:70672041);湖北省自然科学基金(编号:2007ABA307);中央高校基本科研业务费(编号:2010MS112)资助

摘　　要：人民币纸币序列号是由纸币的冠号和数字编号组成,犹如居民身份证号,具有唯一性。纸币序列号的自动识别对实现纸币的有效管理以及缩小纸币真伪鉴别范围具有重要的理论意义和实际应用价值。系统提出了一种基于模板匹配的人民币纸币序列号识别方法。在对图像进行预处理的基础上,先根据物理尺寸将纸币分成不同的类别,然后根据各种面值纸币序列号字符的位置和大小,定位序列号字符;采用投影法分割序列号字符;提取字符网格特征,用特征矩阵表示字符;最后采用模板匹配法识别字符。实验结果表明,系统具有较高的识别率和速率,且具有一定的稳定性。The number of RMB paper currency is unique,like resident identification number.Automation identification to the number of RMB paper currency can improve the management of paper currency and reduce the scope of paper currency authentication.The RMB currency number identification method based on template matching method is presented in this system and the identification system is developed.On the basis of the image pretreatment,Firstly the banknotes are divided into different face value classes by their sizes,and then the number characters are located easily in terms of the banknote＇s number location and size.The number characters in the image are cut up by the projection method,then the character grid feature is extracted and the characters are expressed by Eigen matrix.Finally the characters are identified by template matching.The experiment show that the system has high recognition rate and speed rate,and has certain stability.

关 键 词：人民币纸币 字符识别 模板匹配 图像处理 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]