Automatic License Plate Recognition System Algorithm Research

The Intelligent Transportation System(ITS) is the study topic in the front of the transportation realm of world currently. The Vehicle License Plate Recognition(VLPR), which is an important technique within the Intelligent Transportation System. It takes an important role in information collection vehicle monitor system and automatic fare collection system and other fields. The use of the VLPR is becoming more and more widespread now.The vehicle license plate is the clearest, exactest, and exclusive symbol of vehicles. The VLPR system which is a special vision system, which can record the dynamic data of running vehicles automatically, judge and extract the image data of vehicle license plate effectively and also can recognize the character of vehicle license plate precisely.This paper focuses on three aspect of the license plate recognition system, which are license plate localization based on static image, license plate localization based on video and license plate character recognition. Firstly, this paper proposes a method based on plate texture information and projection method of combining the license plate location algorithm, realize positioning license plate precisely. Secondly, this paper presents an improved BP neural network algorithm and realizes the license plate character of identification rapidly, accurately. Finally, this paper proposes an improved frame difference value method of license plate localization technology, the use of this method can accurately locate the plate of the video.