The Analysis And Research On The Technology Of Detection And Tracking Of Vehicles In Video Sequence

Traffic realtime detection and information collection has become an important subject in ITS(Intelligent Transportation System), and moving vehicle realtime detection and tracking and classification is the capital part. In recent years, the technology on detection and tracking based on video sequence has become the focal point. Through summing up and analyzing the characteristics of existing methods in vehicle detection and tracking and classification, this paper studies these issues including vehicles detection, shadow elimination, vehicle recognition, tracking and classification, which implements an intelligent system of vehicle monitoring for vision-based detection and tracking and classification of vehicles in image sequences of traffic scenes. Moreover, experiments are implemented to demonstrate the validity of practicality of system. The main contents of the study include such aspects as following:(1) Moving vehicle detection: Through analayzing and studying some methods in background extraction and updating backgrougnd, this paper presents a background model based on modified adaptive background method which can extract the realtime no-foreground background image and implement background realtime updating effectively.(2) Moving vehicle shadows elimination: To solve the shadow problem which influences with vehicle tracking and classification, this paper makes a study of gray-level distribution features of shadow and develops an effective shadow elimination method based on analyzing photometric gain properties of shadow and background region.(3) Moving vehicle tracking: Focusing on the problem caused by time-consuming computation and lost objects in tracking due to spliting and merging of vehicles, this paper investigates the basic tracking methods of moving vehicles, and presents a tracking method on finding the maximal weight graph to robustly handle spliting and merging of multiple vehicles and implements realtime tracking.(4) This preliminary system is implemented by C# on Microsoft.NET framework which can complete dectection and realtime tracking and classification statistics of vehicles. Experimental results from real traffic scenes prove that the system is capable of achieving realtime high-accuracy tracking and classification results effectively, and it is robust and practical.