Research On Vehicle Tracking Technology Based On Snake Model

The urban traffic has been stuffing enormous pressure with the global economic development and the increasing motor vehicle users.Both developed and developing countries are in the great test of facing an increasingly serious traffic situation.the The past to solve traffic problems is the traditional approach to construction of roads. But,with the population growing,urban per capita living space to reduce the growing makes less and less space for the construction of roads.At the same time,the transport system is a complex integrated system,only considering the use of high-tech condiering both the road and vehicles together means to solve traffic problems.Vision-based vehicle detection and tracking technology is the key to intelligent transportation system technology.In this paper,major ones are as follows:(1)By in-depth study of Greedy Snake algorithm.A result of not obtaining an ideal coutour by the shadow of building block,the Snake curve still shows accurate Convergence by changing the direction of adding new parameters to control energy and changing in control of the energy coefficient of symbols,adjusting the weight parameters and distribute dynamic Snake point.(2)By comparing advantages and disadvantages among Snake algorithm by Kass, Greedy Snake algorithm by Williimas and Shah and GVF Snake algorithm,I choose the Greedy Snake model as a template.After using the Ghough transform on the template which has been finished an affine transformation,it will choose the best match template as the the initial contour of Snake model by which has a problem of manual positioning.It also makes the Snake algorithm more convenient and automatic in practical application.It obtains a more smoothing curve by taking the Gaussian Pyramid methods in pretreating the video image.(3)It finally achive the vehicle tracking by connecting with the improved Greedy Snake algorithm,the Ghough transform,the pretreating on video image and the Kalman Fiter in the case of shadow by building block.At the same time,it prove the feacibility and the practicality of the improved Snake algorithm by comparing with the intial Greedy Snake algotithm.And carried out only after the algorithm to improve vehicle tracking;by the original greedy algorithm and Kalman filtering for vehicle tracking integration;as well as improved algorithms and Kalman filtering for vehicle tracking combined with the contrast test,which concluded that the election this article The program is more accurate,and practical application is feasible.