Research On Algorithms Of License Plate Location And Character Segmentation In Complex Background

License Plate Recognition System is one of the most important part of the ITS, and it can be applied to various vehicle management to improve managing efficiency and level. Its wide application will contribute to the process of the traffic management automation of our country. In License Plate Recognition System, plate location and character segmentation are two key steps.Because of the complex of image background, the uncertainty of plate position and image quality, many methods of license plate location and character segmentation are not satisfied. Therefore, the research on the algorithms of license plate location and character segmentation is always the hotspot problem. In this paper we discussed the two parts of License Plate Recognition System:license plate location and character segmentation,then made a deep research about some important and key algorithms. The main content of research is as follow:Research on the algorithm of license plate location. This part put forward a method of plate location based on color evaluation function, which is successful applied to the plate location in complex background, and used the grid idea according to the ratio of the plate area to whole image area, which can quickly weed out the non-plate area. Then use color evaluation function again to find a more accurate location in it.In order to improve the result of character segmentation, license plate picture after location should be corrected. This paper use a method of plate correction based on Hough Transform and Characters Projection, and received good results.Research on the algorithm of character segmentation. This part present an algorithm to segment character based on standard position template. In this algorithm, the standard position template is established first, and then the center position of the characters is located by the method of slip match, in order to achieve accurate segmentation.Generally, the images of complex background and license plate skew are used in our experiment. The experiment results shows these algorithms of License plate location and character segmentation in this paper with good antinterference has a lesser limit in background, furthermore location and segmentation are accurate. In sample experiment, the plate location's correction ratio reaches 97% and the character segmentation's correction ratio can reach 98.7%.