Research On The Algorithm Of License Plate Location

With the increasingly high demand of modern traffic system on trafficmanagement system, intelligent transportation system has got more attention. Licenseplate recognition system is an important application of pattern recognition and digitalimage processing technology in the field of intelligent transportation system, and hasbeen widely used in road toll system and traffic monitoring system etc. In practicalapplication, sophisticated filming environment, such as the uneven illumination,camera motion, image acquisition noise, license plate polluted etc, will have a greatinfluence on the recognition results, make images blurred, and seriously affects theapplication of license recognition technology. The influence of these factors mainlyconcentrates in the license plate location, therefore, license plate result will directlyaffects the system function.This paper collects and analyzes the domestic and foreign achievement of scholarsabout license plate location algorithm in recent years. The algorithm for vehiclelicense plate location is studied deeply. The main work is to improve a method oflicense plate location based on mathematical morphology and binary image, whichuses the binary morphology filling to obtain a plurality of candidate license plate area,statistics the number of black and white scanning hopping in the middle line of eachcandidate area to determine the license plate area. This method adopts the one-linescanning, reduces the computation time, and improves the speed of the license platelocation. Finally, a method for detecting the tilt angle of the license plate is proposed.The algorithm is programmed, simulated and validated in Visual C++6.0. Theexperiments show that the proposed method of license plate location satisfies therequirement of accuracy and real-time.