基于无线网络车牌识别系统识别算法的研究

摘要

字符识别是智能车牌识别系统研究的核心问题,针对有线网络车牌识别系统的缺点,将BP算法应用在基于无线网络的车牌识别系统上,重点讨论了用改进的BP神经网络方法对车牌字符的识别;根据我国最新机动车号牌GA36-2007标准,设计出了用于车牌识别的3个神经网络,用MATLAB完成了对车牌字母部分识别的模拟,采用隐层节点22,动量因子0.8进行训练,最后得出实验结果,达到目标误差小于0.001的要求;证明这种方法应用在基于无线网络车牌识别系统上是高效的.%Focused on the improved BP neural network character recognition on the license plate. According to our latest standard vehicle number plate GA36 -2007, designed for the three neural network license plate recognition, MATLAB completed a license plate with the letters part of the simulated recognition. With 22 hidden nodes, momentum factor of 0. 8 for training. Finally experimental results obtained. Achieve the target error is less than the requirements of 0. 001, Show that the method used in wireless network-based license plate recognition system is efficient.