Moving vehicles segmentation based on Bayesian framework for Gaussian motion model

Accurate moving vehicles segmentation is an interesting yet difficult problem in intelligent transportation system. Robust segmentation is obstructed by some problems: the interferential moving objects in dynamic scenes, the vibration of cameras and the change of lighting conditions are all examples. In this paper, we propose Gaussian motion model for moving vehicles segmentation in the dynamic scenes. By investigating the distinction between the motion vectors of the dynamic background and those of the moving vehicles, we find the fact that the motion vectors of the moving vehicles cluster in a small region while those of the dynamic background are dispersive. Gaussian motion model is then proposed to model the motion of the moving vehicles and that of the dynamic background. Bayesian framework is utilized to classify the motion in the scenes to improve the robustness of the model and EM algorithm is used to estimate the parameters of the model. Experimental results on typical scenes, such as waving trees, camera vibration, snowfall and fog, show that the proposed model can segment the moving vehicles correctly in the complex scenes. Quantitative evaluations demonstrate that the proposed method outperform the existing methods.