License Plate Recognition System Designed For Application

License plate recognition system is an important component of intelligent transportation, which plays a critical role in automatic charges of highway and traffic management. PLR has became one of important research subjects in intelligent transportation field that apply computer vision and pattern recognition technology, recently,. With the popularity of the license plate recognition system, in fixed environment and simple scene, license plate localization algorithm before can’t meet the needs of the present. Therefore, license plate identification under complex environment gradually becomes a research focus. This paper mainly focuses on the license plate recognition research in the noise environment.This paper studies noise model and image denoising method, and concentrates on license plate identification research of noise environment. Not only in the overall design to deal with the noise of image, but also in the each link of system design to consider the effect of noise. First of all, image is input and done preprocessing including adaptive denoising,to enable it be transformed into right image that fit for license plate location and identification of subsequent image; Secondly,the image is made an early location based on edge extraction and morphological processing, and made license plate segmentation based on the color. When making edge extraction, comparing the classic edge extraction operators and the sensitivity of noise of them, to select the sobel operator that is not sensitive to noise and can effectively extract the edge of the license plate;Thirdly, through the above methods, obtains the image of license plate, after it making tilt correction and adaptive binarization, to segment characters of license plate with moving window projection method, in this way can effectively reduce the noise influence. Finally, use correlation matching method with stronger robustness to do character recognition for the image.This paper uses MATLAB doing experiments under all kinds of single strong noise and mixed strong noise. Experimental results show that the system of the design can achieve high recognition rate and effectively reduce strong noise interference, has stronger robustness, can meet the needs of license plate recognition under certain noise environment.