Practical Methods for Camera Planning

Abstract

Any multi-camera system requires placement decisions for its cameras to achieve the best system performance. Automated planning algorithms attempt to find the best placement and orientation of the cameras. We propose an optimization based approach maximizing the combined visible volume in the scene. Calculating the visible volume is a computationaly intensive process. We use GPU's to make it practical. Some of the well-known optimization methods (stochastic gradient descent, particle swarm optimization, artificial bee colony and their variations) are implemented and tested on real world scenarios. Experimental results show that the proposed approach can be used in practical applications.