Vehicle License Plate Recognition Technology Research

With the development of the economic , the number of vehicle rapidly increased and the problems of vehicle are prominent more and more. These all make the intelligent traffic system (ITS) is a popular research field and be given prominence to increasingly. The recognition of vehicle license is one of the important questions for discussion of the application in the field of intelligent traffic of the technology of model recognition and the vision of computer. It is a important part of ITS and is applied to dealing with traffic accident, monitoring peccancy vehicles, managing the vehicles of port , aerodrome etc. It has a wonderful future.LPR include three important technology which are the location of vehicle license, the segment of the characters of the vehicle license and the recognition of the character of the vehicle license. There are many ways of recognizing the characters of vehicle license, but the recognition of vehicle license always is a difficult question when the characters are blur, inclined, broken, distorted and smut.In this paper, the algorithm of filter for color image and the arithmetic for the recognition of characters of vehicle license were researched according to comparing with the former recognition technique and integrating the knowledge of image processing and pattern recognition.A fast vector median filtering algorithm for color image is proposed in the paper. The scalar of computation of traditional vector median filter rapidly increased when the window of filtering increased and the speed of the algorithm was influenced. In this paper, direction region distance measure is defined based on the region attribute of pels and the window is divided two children windows according to it, then do the vector median filtering use the pels of the children window. This not only preserve the details of the color image and has a better result of vision filtering but also decreased the scalar of computation of the algorithm and increased the speed of the computing.A algorithm of the recognition of the characters of the vehicle license based on edge hausdorff distance is proposed. The algorithm uses two steps template matching. First, match the characters based on the foreground match measure, background match measure and the whole mismatch measure respectively. Then, do edge detection operations to the mismatched characters. And finally, match the characters using edge Hausdorff distance. The experimental results show the effectiveness of the algorithm. It not only can identify the similar characters, but also offers robustness to the dirty, broken, distorted characters.