Research Of License Plate Recognition Algorithm

License plate recognition (LPR) is one of the pivotal techniques of intelligent traffic system (ITS). In recent years, there has a lot of research on LPR, and many license plate recognition algorithms have been proposed and used. LPR plays an important role in numerous applications such as unattended parking lots, security control of restricted areas, automatic toll collection, electronic payment systems (toll payment and parking fee payment), and freeway and arterial management systems for traffic surveillance and so on. License plate recognition algorithms are generally composed of the following four processing steps:1) image process technologies; 2) extraction of license plate region; 3) segmentation of the plate characters; and 4) recognition of each character. In a LPR system, the input is an image with license plates acquired by digital camera or CCD, and the output is the characters on the license plate. While the license plate recognition rate relies heavily on recognition techniques, sensing technology is essential to the implementation of this process as a system. Most of the LPR systems are used in outdoor environment; therefore the plates may get dirty, be bent and scratched because of lighting conditions, surrounding environmental complexity, vehicle speed and vehicle vibration, while all of these factors add the difficulty of the LPR. In addition, LPR algorithms should operate fast enough to fulfill the needs of ITS.Aiming at the requirement of vehicle management in Changchun and after the analysis of the development and status of LPR, the algorithms of location, segmentation and recognition were deeply researched, this paper focused on the algorithms of LPR based on the automatically taking picture, videoing and saving.Firstly, this system used the gradation image processing to finish the localization and the recognition. Therefore, first the system transformed the color image to the gradation image with the weighting even averaging method, and removed the image noise and the disturbance factor with the middle value filter. To separate the aim from background and protrude the license plate edge, edge detection was adopted. In order to reduce the background picture elements' disturbance, but preserves or strengthens the picture element in the target area, the image should be carried on binarization processing, this article used the algorithm combined Laplace operator and iterative method. License plate location is the fundamental step of LPR, whether the localization is accurate or not is directly related to success or failure of the system recognition. Take the number of edge points of each line in the license plate area and the ratio of the number of edge points and the length of license plate area into count, first texture feature was used to seek approximately localization, obtaining some areas that may contain the license plate; then calculated the aspect ration and eliminated the pseudo areas.License plate character segmentation directly effects character recognition, firstly the correction for grade pretreatment of located license plate should be finished before license plate character segmentation. This method made full use of the width and length of the license plate, character features and other information. Normalization and thinning process were taken place in order to facilitate the character recognition.License plate character recognition is the most important link of the overall system, the recognition methods quality immediately influence overall recognition result. This paper uses characteristic selection based on the statistics and the method which unifies characteristic selection based on character structure to finish the recognition to the license plate character. Firstly, grid characters were used for rough sort; then interval structure characters were used to distinguish the similar characters. The mean and standard deviation were used to describe these target samples. For the characters that the difference is smaller, this paper built more model and template library. According to the position of each character, character classifier was designed and fuzzy classifier was proposed based on the fuzzy decision-making according to above characteristics, the classifier can finish recognition to the single character.On the above work achievement foundation, the license plate localization and the recognition system software that can collect license plate image pretreatment, the localization division, the character segmentation and the character recognition function has developed by using VC++6.0. This software has such characteristics as modulation, visualization, simplification. Finally, related tests were developed to verify the efficiency and realtime.