Bi-objective Optimal Camera Placement with Camera Occlusion Constraints

Optimal camera placement consists of selecting positions for cameras in an environment in order to guarantee a certain level of required coverage whilst minimizing camera costs. In this paper, we delve into this realm focusing on indoor spaces with various types of occlusions, including camera-camera occlusion constraints. We model the problem using both single objective and a bi-objective approaches exploring the total cameras versus coverage tradeoff, solving them exactly via integer programming and the ϵ-constraint method, respectively. Our experiments show that while both models are able to solve the majority of the instances, the bi-objective method leads to significant reductions in the total number of cameras with only mild decrease in coverage, with this approach also being more stable with respect to infeasibilities in environments with high obstacle density.