Research For License Plate Recognition System Based On Linear CCD Camera

License plate recognition (LPR) is a key technique in the modern intelligenttransportation system (ITS) in which the license plate is the input for recognition. Inthe system the license plate can be extracted from the original image and thecharacters in the plate can be recognized in real time. It applies the techniques suchas image recognition and computer vision to process the obtained license plate, andobtain the accurate characters in the plate in real time. The recognition result will bedisplayed with the simple data form of the characters, which make the constructionand management of the license database realizable. As key part in the ITS, therecognition accuracy and real-time of the license plate recognition system willdefinitely affect the progress of the ITS. In this paper, by adopting sidelobe blankingsystem in order to improve anti-jamming capability of the phased array, the mainchannel and guard channel adaptive beamforming in this system is studiedintensively.In this paper, we first summarize the disadvantages in various type of existinglicense plate system, and propose a license plate acquisition system for the shootingconditions of great depression angle. The system aims to acquire the image of themoving car at a high speed in specific traffic lane. In this paper, we study the relatedtechniques and design a license plate system based on linear CCD camera. In thesystem, we scan the image by the linear CCD camera, and take the FPGA as the core.Compared with the common CCD-array camera system, the proposed method canobtain the image with higher resolution for subsequent processing while it cancontrol the gain and film exposure so as to obtain the sharp image in various lightingconditions.Secondly, we present the hardware and software scheme of the system,including the software design of FPGA configuration, USB2.0firmware program andimage buffer. We also study the software solution to the compensation of vehiclespeed variation based on the obtained image.Finally, we propose a license plate location algorithm based on grayscalefrequency variation based on the properties of the acquired image and the frequencyproperty of the grayscale for the characters in the license plate. In the algorithm we obtain the location of the license plate by accumulating the flag value of thefrequency variation of the neighbor grayscale.Results for the shooting experiment demonstrate that the license plateacquisition system based on linear CCD camera is feasible and offer high-resolutionimages for the subsequent processing. The proposed license plate location algorithmcan effectively improve the accuracy of license plate recognition, and the softwaresolution to the compensation of vehicle speed variation has good result.