Research On Several Key Techniques Of Intelligent Vehicle License Plate Recognition

Through the research of four key links of the license plate location, imagepreprocessing, character segmentation and character recognition on the license plate recognition system. This paper designed an intact license plate recognition process and algorithm, and has been simulated in MATLAB.In the part of license plate location, a method based on color character and vein character was adopted to realize accurate location of the license plate. In the part of image preprocessing, gray level transformation, tilt correction and binaryzation were processed to plate image after location. The license plate frame and rivets were removed in this part too. In the part of character segmentation, a segmentation algorithm combined with template matching, vertical projection and clustering analysis was put forward and adopted according to the current license plate design principle, and license plate image was divided into 7 characters waiting to be recognized according to this algorithm. This method is effective on solving the problems of disconnection of Chinese characters, character cementation and noise interference. In the part of character recognition, the improved new method of template matching was adopted, which combined ordinary template matching method and the edge extraction template method.Experimental results show that the new method has good effect on the recognition ofsimilar characters and license plate image which includes faded, dirt or blurry licenseimages.