The Reasearch On License Plate Location Based On Regional Synthetic Characteristic And Its Application

As one part of Intelligent Traffic Management System, vehicle license plate recognition system has a widespread range of application in the Automatic toll collection system at the bridge crossing, parking automatic management system and automatic vehicle violation recording system, etc. The accuracy of the license plate location is one important factor which affects the identification accuracy of Automatic Vehicle Recognition System. Therefore, the license plate location algorithm for segmentation has become a hot issue for research scholars from various countries. In this paper, we aim at a more efficient algorithm to solve this key problem.In this dissertation, the algorithm of license plate location is studied and applied on the electronic police system which relied on the developed work in the Intelligent Transportation System (ITS) project of one city in Hunan Province.Collation and summary of recent advances in the field of license plate location is done, and some of the progress of the latest research fruits is also collected in this collection. After extensive analysis and comparison, this paper presents a license plate location method based on regional synthetic characteristics. This method uses the color feature, texture characteristics, and geometrical features of the license plate, which adopt a more favorable color segmentation of the HSV color space. According to the background and character color plate with a fixed combination of features, we use a simple and practical algorithm for color image segmentation. Firstly, we convert the RGB images to HSV images, and then background color filtering analysis based on HSV color model is done. This will convert the HSV images to five grayscale images. We use morphological smoothing method to decrease the noise of the five grayscale image again. We can get several candidate areas of the plate. At last, by calculating the density, frequency, aspect ratio to separate out the license plate region. This method uses comprehensive feature for location, overcome the shortcomings and limitations of a single feature location, and less restrictive to plate size, location and illumination conditions. The experimental results show that the method is highly adaptable and effective with high accuracy, and has a widespread use in other applications.Based on the above-mentioned study, we designed and realized a license plate recognition system. This system is used in the electric police system's background management software in one city in Hunan Province, which realized license plate recognition and data entry to the database .