Under The Background Of Complex Fast License Plate Recognition Technology Research

Intelligent Transport System (ITS) is an information technology which improves traffic condition by high technology. This technology can make the traffic more clear, safer,and more green.License Plate Recognition (LPR) technology is one of the most important parts in ITS, including License Plate Location, Characters Segmentation and Characters Recognition. As security video has entered the high-definition era and video resolution is more and more high, intelligent transportation system has higher requirements for license plate recognition technology:faster processing speed, greater adaptability to environment and higher recognition rate.Based on the analysis of various domestic and international license plate recognition methods, a comprehensive study in main technical parts of vehicle license plate recognition system has been done in this paper.License plate location:This aspect is an extremely important step in this system. Several algorithms proposed traditionally are compared, then a new location algorithm based on character edge is put forward. The location process consists of two stages:coarse location and exact location. The rapid and simple method is used in coarse location, and the complex and accurate method is used in exact location.Character segmentation:There is a problem that it is not accurate to deal with character adhesion using simple vertical projection. This paper has proposed a method based on the contour of the character and the template matching. This method utilizes character contour features to search and confirm possible characters around contour. Then the found characters are matched according to the adaptive template. The characters are segmented once more by the little error’s template. According to the best template, complement residual characters and remove fake characters.Character recognition:After comparing and testing many kinds of character features, this paper selected effective statistical features and effective structural features as character features. Because SVM classifier has outstanding classification results for a small number of samples, so the SVM classifier is used to recognize the characters. Special features which can distinguish similar characters are used in second recognition of similar characters. Then the license plate recognition system has a high recognition accuracy rate.3000license plate pictures under different light are tested, then the test result:the veracity of palte location is98.6%, recognition rate of last6characters is95.5%, and the average processing time is100ms. The result shows that this license plate recognition method is fast and accurate, and has a strong complicated environmental adaptability and practicability.