摘要
研究人民币智能鉴伪问题,人民币序列号是人民币的"身份证",优化识别人民币的真伪将直接影响经济秩序。由于人民币在流通过程中容易出现磨损、污染和缺损等,导致人民币序列号识别的正确率不高。为了提高人民币序列号识别的正确率,更好地鉴别出人民币的真伪,提出根据BP神经网络对人民币序列号识别算法。首先对人民币图像进行预处理,消除一些不利信息和噪声,然后对人民币序列号进行分割和归一化,并提取字符的特征向量,最后采用BP神经网络对字符进行仿真。仿真结果表明,BP神经网络识别算法的人民币序列号识别正确率达到97%以上,并且识别速度快,为人民币识别提供了有效方法。
Research RMB intelligent identication problem.The serial number of RMB is its ＂identification card＂,but in circulation process,RMB is prone to be worn and polluted,so the identification accuracy of RMB serial number is not high.In order to improve the accuracy of RMB serial number identification,a RMB serial number identification algorithm was proposed based on BP neural network.Firstly,RMB images were pretreated and eliminated to erase the noise information,then RMB serial number was segmentation and normalized.Then the character characteristic vector was extracted,and BP neural network was used to recognize the characters.Simulation results show that the recognition rate of BP neural network can reach 97%,and the identifying time is short,which can meet the requirements of RMB serial number identification very well.
出处
《计算机仿真》
CSCD
北大核心
2011年第12期376-379,共4页
Computer Simulation
关键词
人民币
序列号识别
神经网络
特征提取
RMB
Serial numbers identification
Neural network
Feature extraction
