An efficient framework for detecting moving objects and structural lanes in video based surveillance systems

In this paper, we present a video based detection framework for Intelligent Transportation System. We combine the invariant color features with GMM background model to eliminate shadows and highlighted background regions during moving object segmentation. Then we provide a novel global motion detector to detect camera motions. Finally, we introduce a lane detection technique for automatic structural road detection. Experiments shows that our proposed framework can provide well inputs for vehicle tracking and traffic information detection systems and has computation cost.