The Research Of Vehicle Plate Character Recognition Technology

As a main part of intelligent transport system (ITS), license plate recognition (LPR) has great application prospective in monitoring traffic flow, tolling, investigating on-the-spot traffic accident and handling breaks of the traffic regulations. People pay increasing attention to LPR based on digital image processing and pattern recognition. Automatic License Plate Recognition System is gradually developing into an intelligent transportation system research focus.License plate recognition technology is a wide range of applications technology, Vehicle license plates location, Vehicle license plate character segmentation and vehicle license plate character recognition are three key steps of license plate recognition. License plate recognition technology is studied in the thesis from the three steps. In this thesis, we have been studied on the license plate recognition technology in the license plate character recognition character segmentation and registration deeply. The main contributions of this thesis are as follows:1. Through the research of character segmentation, we propose a method of license plate character segmentation ---- Integrated character segmentation. After median filter and remove the license plate border and rivets, we use the integrated character segmentation method for license plate character segmentation, and get a good segmentation.2. For the license plate character recognition problem, by the study of license plate recognition based on the commonly used methods, Use the determine conditions of symmetric projection In the license plate feature extraction and character classify process. And the method of BP neural network will be applied to license plate character recognition, raised in the traditional methods to improve the introduction of momentum and adaptive learning rate, increase the license plate character recognition accuracy.This algorithm was achieved in Visual C plus plus 6.0 programming environment. Through the experimental showed that for easily confused characters, such as"0"with"D","8"with"B","7"with"T"and others have a better recognition results, lead the recognition accuracy of license plate character recognition have been improved to about 96 percent., recognition speed of the license plate reduced to 0.65 seconds per plate. This method has improved accuracy of License Plate and shortened the time to identify simultaneously.