A Study Of Technology In License Plate Recognition Based On The Radial Basis Function Neural Networks

Vehicle license plate recognition (LPR) is the core technology in the field of intelligent transportation system (ITS), it has very important practice value, the application of LPR will improve the traffic move and management efficiency. LPR system consists of three modules in general, those are: license plate locating, character segmentation and character recognition.Locate the Plate is the first step of the LPR system; its accuracy directly affects the success of the character recognition result. Firstly, image preprocessing is adopted, which is then followed by mathematical morphology operations, secondly, some preparative areas are obtained, and finally, the real license plate area is obtained according to the characteristic of the license plate.The result of character segmentation affects the recognition directly, firstly, slant-correction and binarization are processed, then, a line-scanning method based on transcendent knowledge of plate to is taken to wipe off the frames and rivets. In the paper, the method of characters segmentation is connected vertical projection information and the transcendent knowledge, and then improves the algorithm, which improves the accuracy of segmentation effectively.Character recognition is the very core of the LPR system. In the thesis we structure three networks to apply to characters of LPR based on Radial Basis Function Neural Networks (RBFNN).Through the further training method, the recognition capability of the RBFNN is improved much more.As the test result, it proves that this recognition system can relatively accurate to locate license plate and recognize characters and system's performance is good. It shows that combined pretreatments and recognition techniques can improve the ability of recognition.