Vehicle License Positioning And Tracking System Based On Surf

In the intelligent transportation system, the prerequisite and foundation to achieve the goal of automatic and intelligent transport management is how to make detection in real-time and tracking effectively. Video-based vehicle license positioning and tracking technology in this paper plays an important role in intelligent transportation system. As in the practical scenarios, vehicle license plate detecting system commonly works outdoor, in which circumstances classic algorithm is prone to drift even failure in tracking due to complicated background, various weather, license plate’s oblique and abrasion. I make several improvements and enhancements in this paper for the insignificance of classic algorithm, propose a brand new vehicle license plate detection and tracking method applies to cities’main road.For detecting vehicle license plate, license plate detection algorithm is proposed based on the combination of color and texture analysis in this paper. The algorithm uses both HSV color model to selectively transform various color components to different grey level and the hopping characteristic of the projection map of plate area to detect license plate precisely. The tracking algorithm in this paper utilizes both LK optical tracking and surf feature descriptor matching, combines them to reduce cases of drifting and false tracking, and increase the accuracy of tracking significantly. Finally, I design a scheme of vehicle plate’s detection and tracking system, and put forward solutions to specific issues in this system. The experiments show that, my system has the advantages of good versatility, adapting to different background and illumination compared to the one which bases on single feature, is able to locate vehicle license plate area, works well in real-time circumstances, and both detection and tracking results are satisfactory.