The authors present a framework for placing multiple sensors in an optimal viewing configuration for a feature inspection task. The problem domain is such that a single sensor would not provide adequate coverage of one or more targets; multiple sensors are required to obtain more complete coverage. The task of surveying a known, stationary target is performed by a group of cooperating intelligent vision sensors, which use a synchronous negotiation protocol based on improvement of the collective utility. The experimental results show that, for a variety of targets and environments, the system can place cameras in an optimal viewing position as verified by an exhaustive search of the discretized space.
展开▼
