The Reasearch On The License Plate Recognition And It's Application In Electronic Police System

In this dissertation, license plate recognition based on wavelets and neural network is studied and applied on the electronic police system which relied on the developed work of Qingdao Hisense Network Technology Co., Ltd, including NKTRDP in the 10th five-year plan: the Key Technology Developing and Demonstration Engineering of ITS.With the rapid economy development in Qingdao city, the number of motor vehicles and drivers has appeared fulminic and speedy increase, transportation flux has a sharp augment accordingly, which makes the city transportation face a bigger and bigger pressure. Especially as the cooperation city of Beijing in the 2008 Olympic Games, settlement of the series of problems transportation faced becomes no delay. Therefore, how to control traffic effectively and supervise transportation with modern high technology has an extremely important significance.This paper based on the electronic police system, under the study on the key technology of the electronic police system - license plate recognition, a new fast license plate positioning arithmetic is developed through the wavelets analysis. This method can rapidly position the license plate region by disposing the high frequency image created by wavelets operation. In this paper, an improved BP arithmetic is applied on the character recognition, overcoming the traditional BP arithmetic shortcoming, e.g. training slowly, tramping the region minimum points. At the same time, its simulation and test with C++ language is given. Finally, actualizes the arithmetic with dynamic link library technology.The practical engineering test indicates the arithmetic can realize automatic recognition of license plate fleetly and accurately on the premise of real time and reliability of this system.