Research On Vehicle License Plates Recognition System Based On Digital Image Processing

As a key technology for the road transport systems of the 21st century, ITS’s(Intelligent Transportation System) research, development, and practical application efforts are being made in many countries around the world. As an important content of ITS, LPR’s (License Plate recognition) system can be widely used in urban traffic management, expressway vehicle monitor, electronic tolling, parking lots management, vehicle tracking and community intelligence management etc. License Plate recognition technique, which is based on theory of computer image processing and pattern recognition, is the key part of LPR system. So it is significant to study the technique not only in theory field but in practical application.In general, LPRS can mainly divide into three modules, including license plates locating, character segmentations and character recognitions. Techniques of all key modules related to LPRS are deeply studied and analyzed in this Paper. Finally, it settled down a series of algorithms used in LPRS based on analyzing the typical license plate recognition algorithms in recent years. In the plates positioning, first of all, the edge of the images could be gotten by a series of digital image processing algorithms, and then deal with them using mathematical morphology, forming more connected regions which are in accordance with vehicle features. After that, according to the shape characteristic of plates respectively, we can get connected regions analysis and exclusion, and thus obtain the rough suspected plate areas. In the character segmentation stage, this paper which mainly associated with the information of projection and the priori knowledge make the feature information of character segmentations into the extraction conditions, and achieve the precise positioning of the suspected license plate regions. At the same time, segmentation of the characters could be also realized. The advantages of the license plates location and character segmentation method which are combined include filtering the colors near the plate region, texture and plate characteristics which are not consistent with the interference texture, such as the grid, borders and so on. To a single character recognition problem, a character coding system formulated according to the structural features of characters is presented, using codes to quickly and accurately identify the license plate characters.The testes of the LPR using the proposed algorithms have been conducted. The experimental results have showed that the proposed approaches are reasonable, and they are feasible and effective in application.