Target acquisition in resource constrained stationary camera systems

Involving humans to monitor complex areas using a network of cameras is expensive and prone to errors and fatigue. Hence arises the need for unmanned, cooperative camera systems that can identify targets with intelligent camera motions. Acquiring the targets requires panning of the cameras in a manner that minimizes the time and the contiguous space that goes unmonitored. Average Linear Uncovered Length (ALUL) quantifies the effect of a specific unmonitored area by computing the average length of all lines drawn through the area. This paper considers ALUL as a metric that indicates the target acquisition ability of a particular system configuration. When used in conjunction with temporal constraints, we can intelligently automate camera coverage to improve target acquisition and tracking performance of the system. The results from the experiments confirm that the coverage in constrained environments when the existing camera configuration cannot view large portions of the region of our interest improves when ALUL is considered.