Research And Design Of License Plate Extraction And Recognition System

With the development of soeiety-econolnic, the number of vehicles increases fiercely. Vehicle License Plate Recognition (LPR) plays an important role in Intelligent Traffic System (ITS). It is a combination of the computer vision technology and the pattern recognition technology. Automatic LPR systems can be used broadly in such practical traffic systems as toll gates of highways and bridges, parking lots supervision, urban traffic monitor systems, ports and airdromes.A LPR system generally includes four parts:image preprocessing, license plate location, characters segment and characters recognition. Based on these key technologies, this paper has a deeply research to these four parts:(1) In the image pre-processing, this paper converts color images to grayscale, then enhance contrast to improve the accuracy of license plate location through gray stretch.(2) In the location and extraction of the license plate, with the full use of license plate texture and contrast characteristics of the region, this paper use the algorithm based on two-dimensional image energy to locate license plate, then extract the plate region through the improved color space distance algorithm.(3) In the lean correction of license plate image, this paper use a correction algorithms based on the center of gravity which can avoid the shortcomings of common correction method. License plate character segmentation using a simple algorithm based on fast projection.(4) In the license plate recognition, this paper recognizes the character images through extracting the wide-girdding characters. For the similarity of some characters, extract the difference of the two characters and distinguish them by the two-level neural network.Experiments show that this method is simple, high accuracy, robustness, and it meets the requirements of the real-time system. It can be applied to vehicle control, road pricing, parking management.