Intersection Traffic Flow Counting Based on Hybrid Regression Model

In this paper, we present a new method for vehicle counting using surveillance video used on the road. As the video passing through the road tends to be jammed and blocked, moving slowly and highly regularly, we propose a counting algorithm based on the regression model. We first process the surveillance video to extract features from the video. Then we use the SVM classifier to verify the vehicle density. Finally, we tested our approach using real pass-through video. Our experiments show that this method can count the flow with higher accuracy and lower computational cost.