License plate recognition system

This paper proposes a license plate recognition system, where the neural network concept to be applied. The identification of license plate is on the basis of character. In order to make the system quickly and accurately the proposed system has designed a corresponding workflow and complete through the PC. When using license plate recognition systems, photographic lens will return successive still images. The system includes a main algorithm before image processing, detection and capture license plate location, license plate cutting element, license plate element feature extraction, as well as license plate recognition and other components several parts. When looking at the license plate, count the number of color changes in the image of each horizontal line. Through the qualifying threshold selected area can discover out the license plate placement. The initial size of the region of localization results will serve as the basis to judge detect abnormal number plates, conditional iteration binarization, modified, and research. Whereby the plate to achieve better positions, but also enhance the accuracy of the cutting plate element. In the plate element cutting part, we use comparative characteristics of the target element and background to detect the color change, to identify elements borders and cutting.