Dempster-Shafer based multi-view occupancy maps

We present a novel method for calculating occupancy maps with a set of calibrated and synchronised cameras. In particular, we propose Dempster-Shafer based fusion of the ground occupancies computed from each view. The method yields very accurate occupancy detection results and in terms of concentration of the occupancy evidence around ground truth person positions it outperforms the state-of-the-art probabilistic occupancy map method and fusion by summing.