Research On Vehicle Detection Based On Video Technology

With the rapid development of world economy, the amount of city vehicles becomes larger and lager. How to make an improvement on the traffic state becomes a focus of governments all over the world. Under this kind of background, the technology of ITS(Intelligent Transportation System) based on video develops quickly. The vehicle detection technology based on video plays a great role on traffic surveillance and controlling, and has great impact on work after then directly. Vehicle detection technology has important practical value in this field.This paper presents a detailed analysis of development state of vehicle detection technology, and the difficulties as well as the problems to be solved. It focuses on the background renewal and the blocked vehicle segmentation techniques, and proposes two new algorithms.This paper first briefly introduces the traffic system only, the development situation of the summary of the problems in the present research. Aiming at the problem of fast extracting road background image in complex scenes, an effective algorithm about road background extraction and updating is presented. An improved algorithm for time averaged background images (TABI) is used to the background extraction; and then an improved surrender algorithm is applied to the background updating. Experimental results show the proposed algorithm can reduce the influence of initial static vehicle on the background extraction and eliminate in time the ghosts caused by the motion of objects in the initial frame. The proposed algorithm is robust under lighting variations, fast in speed and satisfactory in updating performance.As blocked vehicle segmentation is a difficult problem of vehicle detection technology, a blocked vehicle-based active contour and watershed segmentation algorithm is proposed. The algorithm uses the active contour model for accurate positioning of moving vehicles, and then has the shadow detected, the watershed algorithm is used for the blocked vehicles’ segmentation. The experimental results show that this algorithm has a good segmentation results to the blocked vehicle.