基于并行模糊神经网络的车牌识别研究  

Study of Vehicles' License Plates Recognition Based on Parallel Fuzzy-Neural Network

作　　者：周开军[1] 陈三宝[1] 

机构地区：[1]武汉理工大学自动化学院,湖北武汉430063

出　　处：《计算机测量与控制》2006年第4期477-478,490,共3页Computer Measurement &Control

摘　　要：车辆牌照的自动识别是目标自动识别的一种重要形式,针对车牌识别的后期技术,即牌照识别技术做了研究并提出了一种新的车牌识别方法,该网络由BP神经网络识别模块和模糊控制器构成,为了便于硬件实现,各模块相互独立,最后利用PVM网络在虚拟并行平台上实现了该识别网络,实验结果表明,和标准BP网络相比,该算法具有更加良好的性能,满足实时识别车牌的要求。Automatic license plate （LP） recognition is an important form in the automatic target recognition. A new method of later technique of LP recognition was studied and proposed, which was constructed by the BP neural network recognition module and fuzzy controller, In order to be implemented conveniently by hardware each module is independent, Utilizing PVM network the recognition network is implemented on the virtual parallel platform in the end, Experiment results showed that the algorithm has nicer performance than standard BP network, and satisfies the requirements of real--time LP recognition,

关 键 词：BP神经网络 并行处理 PVM网络 车牌识别 

分 类 号：TP911.7[自动化与计算机技术]