Traffic Object Tracking Based on Increased-step Motion History Image

A new image-based method for tracking moving vehicles is purposed using the layered step-down grey value silhouette of increased-step motion history image. Real-time tracking of moving vehicles appeared in the traffic scene can be achieved by segmentation and marking the motion silhouette regions in the increased-step motion history images. High quality segmentation is realized by improve the basic motion history image. Background of the traffic scene is subtracted from the traffic video frames. At last, experiments are done with crossroad traffic videos. Moving objects is segmented effectively. The results show that the method is robust against the disturbance of changeful environment, high detecting rate and fast real-time processing speed.