基于神经网络的钞票真假识别研究  

Identification of Counterfeit Banknote Based on Neural Networks

作　　者：杨中东[1] 雷玉堂[1] 

机构地区：[1]武汉乐通光电有限公司高新技术研发中心,武汉430074

出　　处：《光学与光电技术》2006年第5期108-111,共4页Optics & Optoelectronic Technology

摘　　要：利用神经网络与光电检测的技术研制了钞票真假识别系统。介绍了系统的结构组成、工作原理、软件系统、神经网络的优化设计、实验及测试结果。经实践验证,其识别结果稳定可靠,可应用于金融智能防伪点钞机与ATM机中。A counterfeit banknote identification system is developed by using neural network techniques and opto-detection. The architecture, principle, neural network software optimization design, experiment and test results of the identification system are introduced. Confirmed by the practice, identification result is reliable and the system can be applied to the financial smart anti-counterfeiting and banknote counting machine and ATM.

关 键 词：神经网络 神经元 隐含层 权值 训练模式 识别 

分 类 号：TN911.73[电子电信—通信与信息系统]