Vehicles´ Motion Parameters Detection in Intersection Based on Videos of Low View Angle

By taking Visual C++ and OpenCV as tools and processing video image sequences gathered by one camera under low view angle, suitable foreground detection method of intersection was selected. Through improving blob tracking system provided by OpenCV, the program was developed to detect motion parameters of vehicles in intersection. The motion parameters were smoothed by Kalman filter. It made the data more reliably. The research could be applied in detections of the traffic conflict and traffic information.