Research And Application Of Vidio License Plate Recognition System Based On Software Trigger

Recent years, the number of private car owners in our country is increasing,which advances the development of some automobile-related industries and drives the vehicle plate recognition system to be applied in more applications,such as places where it is not convenient to install outside triggering device or mobile checking system,etc.Vehicle plate recognition system mainly uses digital image processing technique to analyze the image, so it can extract the information of the plate and insure its accuracy. Automatic extraction of the plate number rolls vital significance as it has higher efficiency compared to that of artificially processing and exerts irreplaceable role in some ways. License plate identification technology overseas have developed very mature, and many developed countries and regions have designed corresponding plate recognize system according to their own plate's characters. However, as the license plate has its unique characteristic in our country, foreign product does not apply to the actual conditions of our country.In our country, there are also some scholars and research institutions developed similar products. But most of these products are based on numerical calculation, and need external trigger equipment, which causes the system cost too much, and then influences the benefits of users.This subject based on the development of vehicle plate recognition subsystem which is used in the ERP system of 4S stores design a vehicle plate recognition system, and it gains the image of the plate by software to analyze vedio information, which can successfully realize the automation of plate recognition without outside triggering device.The main contents of the subject are as follows:Firstly, according to the implementation of the license plate recognition system, the system determined the way of acquiring images. Then, we introduced the workflow of the license plate recognition system, the main equipments of the system, and the connection mode of the main equipments as well as selection of video acquisition equipment.Secondly, we design the system's software structure on the basis of the workflow and the system's hardware, and chose the Visual C++ programming environment to develop system's software. We use double processes to analyze video and identify the license separately, then discuss the communication method between the double processes, the system's database support, and the reliability design of the system.Thirdly, we discuss the main algorithms of the system based on hardware and software structure. Among them, the vehicle triggered part mainly introduces the realization method of vehicle trigger function ,the way of avoiding mistake triggering and trigger missing; the license plate location part adopts edge detection method combined with color characteristics to complete license plate localization; in the characters segmentation part ,we first optimize the plate image by adopting the image preprocessing algorithm based on prior knowledge, and then complete this part according to image vertical projection; on the character recognition phase, we chose character sliding matching method based on multy-eigenvalue to complete recognizing characters, and discussed the special treatment of confusable characters and improve the system recognition accuracy of confusable characters. All of these algorithms are completed in the Visual C++ programming environment.In the end, we tested the function of the system comprehensively in the system's implementation site and analyzed the result. The result of experiment shows that the algorithms we used in this paper can obtain the vehicles image and the license information precisly, and we achived the expected result.