摘要
分析了反传学习神经网络和Hopfield神经网络的基本原理,探讨了神经网络在汽车牌照字符识别中的应用。结合神经网络和汽车牌照的特点,研究了学习速率,误差精度与隐含层节点数之间的关系,最终提出了一种Hopfield神经网络和反传学习神经网络相结合用于汽车牌照字符识别的方案。Matlab仿真结果表明,所设计的汽车牌照字符识别系统可以获得较为满意的高分辨率。
The basic theories of back-propagation network and Hopfield network are analyzed, and the application of the two networks in the vehicle license character recognition is discussed. With combining neural network and characteristic of vehicle license, the relations among the study rate, error precision and nodes of the hidden layer are studied, and a method of vehicle license character recognition by combining back-propagation network and Hopfield network is put forward. Through Matlab simulation, it shows that the system can get lower misclassification in the vehicle character recognition.
出处
《计算机工程与设计》
CSCD
北大核心
2008年第8期2041-2043,共3页
Computer Engineering and Design
基金
广东省教育厅自然科学研究基金项目(Z03076)
广东省科技计划基金项目(2006B12701002)