License Plate Recognition System License Plate Location And Character Segmentation Method

Vehicle license plate recognition system (VLPR) is one of the important parts of intelligent traffic system (ITS). It can be used in the fields such as electronic toll collection, pass controlling, traffic monitoring and so on. VLPR consists of three main modules which are license plate location, character segmentation and character recognition. The result of the former two modules is the key of success recognition of character. Based on the summary of newest research results in the field of license plate location and character segmentation, deeply research on algorithm of license plate location and character segmentation has been done.In the stage of license plate location, lightness component of the image should be balanced in HSI space at first, so the suitable scope of image lightness has been enhanced, and the color information of image can not been affected. And color pair pixels are extracted in HSI color space according to the rule that character color and plate color should be matched. Mathematics morphological operation which use dynamic operator is adopted in order to extend the color pair in real plate to a whole area. Then, the image is scanned line by line to get the areas that seem like real plate.Slant correction should be done before character segmentation. A horizontal slant correction approach based on color pair pixels area and principal component analysis is presented. Through principal component analysis of the color pair pixels area, the slant angle of the plate can be obtained. When revolving adjustment, filling process should be performed to guarantee the integrity of the image information.In the stage of character segmentation, plate area is located by shape of color pair pixels area, and pretreatment including gray proportion, binarization and edge detection is taken before extract the character area. The most suitable threshold value algorithm has been selected on plates in different colors dring the process of binarization. Then, according to the characteristic of character size proposed that the dynamic template method carries on the character division, and transform characters to identical size.Finally, a software platform is constructed by language of C. This software contains all the steps that mentioned above. Through the experiments on large numbers of pictures especially on which have massive disturbances, inadequate lighting and multi-car licenses, the algorithm of license plate location and character segmentation has been proved to be real-time, steady and good at anti-jamming, which create favorable conditions for the following character recognition.