Research Of Pick-up And Recognition Of Vehicle License Plate

With the increment of vehicle all over the world, the situation of city traffic is attracted the recognition of people. How to manage the traffic effectively is become the focus of people more and more. License Plate Recognition System is one of the most important part of the traffic management system. Through distill the license plate and identify character, License Plate Recognition System accomplish the function of charge automatically, inspect, manage and demand, and so on. The article research license plate recognition, including the collection of picture of vehicle , distill of the license plate, segmentation of character and recognition of character. The main content of research is as follow:Design of the collection system vehicle pictures. The article design a set of collection system in view of the idiographic case and the demand of real-time: the loop induce the vehicle is passing, and give the signal to controller, the controller will control the card of image collection to screen a picture, getting a vehicle picture including license plate.Researching the pick-up algorithm of license plate. First, pretreating the images, including gray the RGB image, linearity transform to gray image, and make to gray image to black and white images, and bring forward the method of combine the quality of projection gray scale and property of license plate, distill the license plate .It is satisfy the demand in speed and veracity.Researching the segmentation algorithm of character. At first emendate the license plate, then utilize the quality of character, such as width of character and space between character, to segment the character. The method is simple and applied, and the result of segmentation is good.Researching the recognition algorithm of character. After research some ordinary methods about recognition of character, bring forward the improved method of template matching. First, make the character to the same scale, and get the framework of the character, then get the stamp of character, comparing the vector of stamp with the vector of the character in the template respectively, the result of the least error as the result of the recognition. Through the method can get higher veracity and faster speed.The algorithm we adopted is not restricted by the location of the plate in the image, and it has strong anti-jamming ability to lean, distortion, pollution and blur. The algorithm is also adaptable to the varieties of circumstance light and image contrast.