The Method Of Locating License Plate Based On Color And Black-white Texture Analysis

With the rapid development of our social economy, the quick increase of vehicles and the more and more busy of traffic transportation, the traffic jams and accidents are becoming the great puzzling problems in our country, especially in the urban. Intelligent Transportation System(ITS) is presenting to and developing in such embarrassing situation.License Plate Recognition(LPR), is an important part in ITS and the key technology in the field of highway traffic automation. It shows a great application value in public safety, traffic control, custom, military and so on, and it is currently becoming a hot research spot of the people all over the world.LPR are mainly formed by three steps, including license plate location, character division and character recognition. The license plate location is the foundation of the license recognition, it plays the most important role to the performance of all system, and it decides the speed and precision of the recognition system. Thus license plate location is required to hold the information in the plate region completely, and delete all other disturbing information out of the plate region as much as possible.The dissertation studies the license plate location technology of color automobile image using digital image processing technology. First it introduces the up-to-date status of domestic and foreign development in license plate location technology and analyzes the difficulties in license plate location technology, then it discusses imagery processing method used commonly, and next it conducts a thorough research to it. According to the main characteristics of license plate -- texture, color and structure, and by taking full advantage of the characteristic of fixed color match between the plate background and plate character in the color image, a method to locate the plate is put forward based on the characteristic of its color, black-white texture and structure. That is: based on the differences between the background color of the plate and the character color, the research is focused on picking up and analyzing the color characteristic in plate region, distilling out the color regions consistent with the plate background color and the plate character color respectively in HSV space, then the plate region is extracted out by synthesizing multiple characteristics of plate texture, structure, and so on.The algorithm in the dissertation has taken into account the characteristics of the texture, background color, character color, structure, and so on, and the operation has been carried out in binary image as much as possible. The result shows that the method can use multiple characteristics properly to locate the plate with fast speed and high accurateness, the application scope is broad and the adaptability is strong, and the algorithm is robust and real time in certain extent.