Research On Algorithm Of License Plate Location Based On Static Vehicle Images

License Plate Recognition (LPR) system is an important part of the Intelligent Transportation System (ITS), and plays an important role in modern traffic management and social security. License Plate Detection is one of the key technologies in LPR systems. With the extensive application of LPR systems, the algorithms that used to work under controlled conditions and simple scenes is unable to fulfill the requirement. Therefore, license plate detection under complex scenes and imaging conditions becoming a research focus. The LPR system consists of three modules in general, those are: License plate location, character segmentation and character recognization. And we mainly focus on the former one part in this paper.In this paper we have summarized the lastest research achievements and development of license plate location, cascade license plate location algorithm is used to locate static vehicle images. In order to solve the problem of license plate localization, through the analysis of characteristics for vehicle license plate images, geometric feature, character tightness feature and Haar feature were found. In order to describe Haarfeature, Discrete AdaBoost (DAB) and Real AdaBoost (DAB) algorithm that belong to AdaBoost family algorithm were introduced to design classifier, and finally achieved the license plate location by using AdaBoost algorithm based on Haar feature. In order to improve the accuracy of classifier, RAB and DAB algorithm is used to the design of classifier, respectively. Through the quantitative and qualitative analysis, and the experimental result indicates that RAB algorithm is better for license plate location.The process of cascade license plate location algorithm is divided into three aspects. Firstly, extracting all possible candidate plate, according to license plate compact character and geometric feature. Speed up the license plate localization algorithm because of the elimination of the many nonlicense plate areas. And then do the operations of geometric correction for the candidate area of big tilt angle. The last step is the identification of candidate license plate, a large number of samples produce the AdaBoost classifier that is based on Haar feature, through the caculation of this classifier, final license plate localization can be gotten. The experimental results indicate that the experimental algorithms in this paper can be adapted to complex background, and they are excellent in accuracy and robustness.