基于形态学和神经网络车牌识别算法.docVIP

基于形态学和神经网络的车牌识别算法 第25卷 第3期 2001年9月 武汉理工大学学报(交通科学与工程版) Journal of Wuhan University of Technology (Transportation Science Engineering) Vol.25 No.3 September 2001 A Recognition of Vehicle plate with Mathematical Morphology and Neural Networks① Yang Jie Li Xu Guo Wei (College of Information Technology,WUT,Wuhan430063) Abstract:Requirement of automatic recognition of vehicle license numbers from their plates is increasing for social services such as traffic control and collection motor-way charges. This paper proposes an algorithm based on mathematical morphology operations to detect the region of a vehicle licence plate from an image. The uniqueness of this algorithm, in comparison with other approaches to this problem, is that there are no special restrictions imposed on the input images. Automatic recognition and classification based on feedward three layers neural networks are used. All algorithms in the paper are verified at 586 computer and experimental results are efficient for recognition and classification. Key words:mathematical morphology;neural networks;recognition 0 Introduction Traffic engineers and the police often wish to able to uniquely identify road vehicles at different points in a road network for reasons, such as: 1) automatic toll-collection, road pricing schemes, and operation of car parks for revenue collection; 2 ) fleet control for monitoring and optimizing the movements of particular types of vehicles; 3 ) speed and weight-limit enforcement purposes; 4) crime detection / prevention; 5) traffic data collection etc. Until quite recently, vehicle identification has largely been a labor-intensive manual activity, aided where possible by electromechanical actuators. With the continuing increase in volume of road traffic, such manual techniques have proved impractical. Improvements in electronic sensor technology have allowed for the development of automatic vehicle identification (AVI) equipment. Conventional AVI systems employ microwave and in