The Research And Development Of License Plate Recognition System(LPRS)

With the high-speed development of our national economy, there are more and more construct in the domestic expressway, urban road, and parking area, etc. The requisition on the traffic control, safety management improves day by day. The intelligent traffic system (ITS) has already become the main direction of present traffic administration development, and vehicle license plate recognition (LPR) system as core of ITS, plays a very important role. It occupies important status in the project management of expressway, urban road and parking area, etc. Its wide application will contribute to the process of the traffic mechanography of our country.Being a special computer vision system in the real-time case, the LPR system mainly includes the subsystem of car flow statistic and car style measure, and the subsystem of car license plate recognition. At present, the subsystem of car flow statistic and car style measure already finished basically. The subsystem of car license plate recognition is composed of license plate detection and character recognition. The LPR system involves numerous discipline domains, such as Pattern Recognition and Artificial Intelligence, Computer Vision, Digital Image Processing etc. Its key techniques include the license plate detection, the license plate image pretreatment, character segment and the design of classifier. The primary direction of my research is the study of license plate character recognition, including the pretreatment of license plate image, license plate character segment and character recognition. The typical difficulty of the system is the stochastic interference of external environment because the license plate images come from natural scene. So I must try to reduce external influence and stress the image feature.The pretreatment of license plate image is composed of reduce influence, image buildup, image binarization, distortion and anti-distortion and so on. Further more the image pretreatment is connected closely with the acquisition of image feature. The LPR system processes the natural image with stochastic disturbance. The paper presents a pretreatment method able to solve the problem that the image contrast degree is too lower to our eyes because of unconstrained illumination conditions.Single character image segment means that the whole license plate image will be divided into several little parts and then followed by character recognition. Each one includes only one character respectively. The difficulties lie hi that there is any disturbance of yawp, character conglutination, and character rupture etc. The paper first presents some methods that adopt prior knowledge of the license plate, perpendicular projection, and minimum character area and so on. The detail is that starts from the fourth character image (the middle of license plate), centers it, and then segments every character seriatim to the direction of left and right. It makes full use of the feature of the image. The priori information of license plate does well withthe character conglutination and character rupture and presents a nice pre-segment. At last the whole license plate is well segmented into seven parts.The method of character recognition takes an important part in the whole system, and it has direct effect on the last license plate recognition result. The key technique lies in the matching of feature and classifier. Typical difficulties are the character blur, character destruction and character image stained because of the performance of the vidicons, the clean degree, illumination conditions and the movement of the vehicle and so on. The paper presents a method using several image features to solve the problem. In the end I get a good license plate recognition result.