| With the rapid increase of vehicle, the command for traffic control and safety management is getting higher and higher, intellectual traffic system has already become the main direction of present traffic administration development. The vehicle license plate, as the only sign of the exercising vehicle, has irreplaceable function in control and management of the vehicle. Based on treatment technology of the picture, the license plate recognition system of vehicle is regarded as an important part of the intellectual traffic, play an important role in such project management as the expressway, urban road, parking area, etc.In the research of Vehicle License Plate Recognition(VLPR) system, we analyze and process the vehicle image combine with the processing of grayscale , image denoising, gradient sharpening , grayscale texture analysis, histogram statistics ,image projection, image feature extraction and image binarization. The thesis bring forward a method of the location of the vehicle license plate that base on grayscale texture ansalyze method processes the vehicle image by use of gradient sharpening,detecting the discontinuity point ,filtering the discontinuity point by the information of the width of character outline in the license plate region, then projecting the image ,we finally locate the position of the plate with the help of the license plate character size and, the effect is fairly well. According to the result of the location of the vehicle license plate , considering of the clearances between the character and between the characters and the edge of width of the character and the width of the license plate is different ,we project the image in the direction of horizon and vertical, wipe off the edge of the license ,segment and normalize the characters. Considering of the character feature in the license plate region , we bring forward a recognition method of vehicle license plate character ,a BP neutral network recognizer is devised . At the same time, we use VC++ as the develop language realize the relate algorithm during the image preprocessing and VLPR. |