Researches On Vehicle Tracking And Traffic Incident Detection Methods Based On Image

With the increasing vehices on the highway, traffic incidents also become more frequently. Traffic incident detection gradually becomes an attrative topic in transportation field. Traffic incident detection system is an important part of traffic monitoring system. This paper mainly focuses on vehicle detection, vehicle tracking and traffic incident detection. The work and achievement basically has the following three aspects:1. Vehicle detection: In vehicle detection process, we improve the traditional background subtraction model, employing RGB three channels subtraction respectively instead of gray subtraction. Thus we can get more vehicle pixels in the foreground detection and avoid vehicle rupture even disappear to some certain extent. In order to eliminate the impact of the shadow, the algorithm of moving cast shadow detection of vehicle using HSI and c1c2c3 models is proposed. In the post processing, it corrects the shadow detection and vehicle detection errors. This algorithm can detect cast shadows of moving vehicle real-time and accurately.2. Vehicle tracking: In order to track the vehicles real-time and stability, this paper proposes the vehicle tracking algorithm based on SIFT characteristics and region matching. We only extract SIFT characteristic in the vehicle area instead of the whole image, reducing the computation time. Combined with region matching tracking information, we can track the vehicles more accurately.3. Traffic incident detection: Based on vehicle tracking, according to the point imformation of vehicle circum-rectangle center, the vehicle movement direction can be calculated using straight line fitting on the base of tracking trajectory. Then we can judge whether vehicle changes the lane, regression or stop. Establishing mapping relationship between image coordinate system and the world coordinate system, vehicle movement speed can be calculated. Through operation regulations of road top speed of vehicle, we can determine whether vehicle is over speeding.Finally we summarize the work of this paper and illustrate the deficiency of this paper. The further research is pointed out at last.