Research And Design Of License Plate Recognition Algorithm In Public Traffic Management System

With the development of intelligent informatization into the modern society,the syst em platform of information management in the city gradually chooses to use the new ge neration of technical means to manage information to achieve more perfect analysis,inte gration and processing capabilities.Public transportation information is an indispensable p art of urban daily life,and as the number of vehicles is rising in the city,a fast and eff icient traffic information management is essential.This paper studies the public transportation information management platform syste m,and combines the existing research results on the direction of license plate recognitio n at home and abroad.Through the related technology of machine learning,this paper re alizes the automatic recognition of the license plate characters of the image related to th e license plate information in the system.effect.This paper analyzes the actual needs of the public transportation information manag ement system.The license plate recognition module algorithm is based on OpenCV(open source computer vision library)cross-platform computer vision library.The license plate recognition process is mainly divided into two parts: license plate location and license pl ate character recognition:License plate location: Analyze and process the image containing the license plate partial block to obtain the license plate block,and use the HSV color positioning model and the Sobel operator edge detection algorithm to combine the license plate positioning.The pre-selected license plate tile is then judged to finally obtain a tile containing only l icense plate information.The license plate determination process is primarily trained by a support vector machine(SVM)algorithm to determine if the obtained partial tile is a re al vehicle license plate image.License plate character recognition: The image obtained after the license plate is po sitioned is divided and recognized by characters,and arranged in order to form ordered character type license plate information.The learning and classification is completed by t he convolutional neural network CNN model and the character recognition process is out put.The main algorithm of the license plate character recognition module is based on t he Open CV open source visualization library,written in C++ as a writing language,and run and tested using Visual Studio 2013.The research results of the vehicle license plat e recognition function in the public transportation information management system can m eet the business requirements.