基于人工神经网络的车牌识别  

Vehicle Plate Recognition Based on Artificial Neural Network

作　　者：吴聪[1] 殷浩[1] 黄中勇 刘罡[1] 

机构地区：[1]湖北工业大学光庭实验室,湖北武汉430068

出　　处：《计算机技术与发展》2016年第12期160-163,168,共5页Computer Technology and Development

基　　金：国家自然科学基金青年项目(61300127);湖北省自然科学基金项目(2012FFB00701)

摘　　要：车牌作为不同车辆的唯一标识,其识别技术是计算机视频图像在车辆牌照识别方面的一种重要应用,在各种场合是识别汽车身份的重要途径。由于现阶段技术的不断提升,识别过程中的问题也不断涌现,而在车牌预处理、分割以及识别阶段中,车牌识别是现代交通系统中非常重要的功能模块,而其关键因素在于汉字、数字以及字母的识别。通过提高车牌的识别率来提高交通部门的工作效率。目前,人工神经网络因其优越性被广泛应用于各种图像识别中,但因其收敛速度慢,运耗时间长,对实际应用产生了很大的限制。采用遗传算法与神经网络相结合的方法并进行了仿真,实验结果表明,该方法对车牌有很好的识别作用,具有时效性和鲁棒性。License plate as a different unique identification of the vehicle, its identification technology is an important application for com- puter video image in the license plates recognition, which is an important approach to identify a car for a wide range of situations. Due to the improving of the technology at present, the problems are also emerging in the process of recognition. In the stage of preprocessing, segmentation and recognition of the license plate, license plate recognition is an important function module in modem traffic system, and its key factor lies in the identification of Chinese characters, numbers and letters. By improving the license plate recognition rate, the working efficiency of the transport sector is improved. At present, the artificial neural network is widely used because of its superiority in all sorts of image recognition,but because of its slow convergence speed and long time consuming, a lot of restrictions are produced in actual applications. By adopting the combination of genetic algorithm and neural network, the simulation results show that the experimental results of the license plate has good recognition effect, which proves that the method is effective and robust.

关 键 词：车牌识别 神经网络 机器学习 遗传算法 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]