The Moving Vehicle Deection And Tracking Based On DSP

With the rapid development of modern information technology, intelligent system isimproving and updating, and the application of intelligent monitoring system isconcerned in many fields, such as military and civilian. To detect and track themoving vehicle has broad prospects in the field of Intelligent Transportation System,which becomes a very important topic.In this paper, active contour model is used to detect and track the moving venicle.The platform is the DSP hardware platform at the core of TMS320VC5509, which ismade in TI Company. First, analyzing the advantages and disadvantages of commonalgorithms and according to the characteristics of video processing and therequirement of system real-time, the active contour model was chosen for the mainalgorithm. However, the traditional active contour model has many shortcomings; forexample, the dependence on initial contour is strong and it can't detect the targetoutline deep concave part, etc. Aiming at these shortcomings, the improved activecontour model - Gradient Vector Flow (GVF) model is used to solve these problems.In the process of vechicles testing, setting the initial model for vechicle firstly andaccording to the astringency of GVF model itself, vechicle's contour was detectedsuccessfully, which can achieve the detecting purpose. Adopting prediction algorithm,the variation of goal contour can be analysised from detected target. And then theposition and shape of goal contour can be predicted, which can achieve the detectingpurpose using the convergent iterative procedure. From the actual trajectory ofvehicle and the tracking trajectory of algorithm in this paper, the algorithm can detectand track the vehicle effectively, which has advantages of good robustness and highreal-time, etc. With C language designing and optimizing the program code, theoptimized code has faster execution speed and higher efficiency, which proved thesystem achieving good effect after analying the whole system.