百元人民币冠字号码提取与识别

Extraction,Identification on Crown Word Number of One Hundred Yuan RMB

对百元人民币冠字号码的提取与识别,是监督管理百元人民币的有效手段。使用特征定位法定位冠字号码在百元纸币中的区域,用投影法对字符进行分割,对分割后的字符进行归一化处理,实现对冠字号码的提取与归一化;采用改进的KNN算法,实现对冠字区、号码区进行分离,然后对字符进行KNN算法识别。通过算法对比可见,采用改进的机器学习KNN识别算法,能够更好提高码字区域的0和1的识别率,从而提高整个冠字号码的识别率,实现对百元纸币冠字号码的识别。本研究提出的方法简单可靠,具有重要实际的应用价值。

Extracting and identifying the RMB Crown Word Number is an effective technical method used for supervision and management about a hundred paper note.The area of RMB one hundred Crown Word Number can be located based on the feature location way.The vertical projection method was carried on to divide into each character.And then the normalization processing is performed.Improved KNN algorithm can be used to identify the character which used KNN method by making a distinction between Crown area and Number area..The comparison of algorithms shows that improved KNN recognition algorithm can better improve the recognition rate of 0 and 1 in the Crown area and Number area,so as to improve the recognition rate of Crown Word Number and realize the recognition of the RMB Crown Word Number.The method in the study is simple and reliable,and will generate an important application value in practice.