Abstract

This paper presents a highway vehicle counting method in compressed domain, aiming at achieving acceptable estimation performance approaching the pixel-domain methods. Such a task essentially is challenging because the available information (e.g. motion vector) to describe vehiclesin videos is quite limited and inaccurate, and the vehicle count in realistic traffic scenes always varies greatly. Totackle this issue, we first develop a batch of low-level features, which can be extracted from the encoding metadata of videos, to mitigate the informational insufficiency of compressed videos. Then we propose a Hierarchical Classification based Regression (HCR) model to estimate the vehicle count from features. HCR hierarchically divides the traffic scenes into different cases according to vehicle density, such that the broad-variation characteristics of traffic scenes canbe better approximated. Finally, we evaluated the proposedmethod on the real highway surveillance videos. The results show that our method is very competitive to the pixeldomain methods, which can reach similar performance along with its lower complexity.