The Research Of Location Algorithm In License Plate Recognition

With the continuous progress of society and leap development of economy, the level of people’s material life is unceasingly improved, and the auto industry also rapidly develops which result in the rapid increase in holding volume of cars. In order to effectively manage the continuously increasing vehicles and relieve the traffic pressure of roads, a kind of technology that can swiftly, accurately and automatically position and recognize license plate in various environment, license plate recognition (for short, LPR), comes into being. It can output the license number information of the license plate image it took through accurate positioning, character segmentation, character recognition, so it is a kind of key technology for recognizing license plate.The complication and uneven illumination of auto image background, which results from various kinds of natural environment, seriously affects the positioning for license plate. Consequently, how to accurately and effectively locate the area of license plate according to the license plate image collected in various kinds of environment becomes the key for license plate recognition technology.On basis of consulting multiple of representative domestic and overseas positioning algorithm in relevant license plate recognition technology, the author has conducted a comparison based on the existing license plate positioning algorithm by reference to the unique characteristic of our country’s license plate. In this paper, the license plate positioning method based on grey image and the method based on color image are introduced, analyzed and compared. The new algorithm of license plate positioning based on texture analysis and HSV color is analyzed from the perspective of practical applicability and verified with MATLAB as the experimental tool. What’s more, that algorithm is of high robustness with the advantages of little limit on image background, good noise acceptable capability and quick positioning.