Correlation model for randomly deployed heterogeneous cameras in wireless multimedia sensor networks

In practical wireless multimedia sensor networks, efficient utilization of available resources is very critical. However, in dense networks with randomly deployed camera nodes, it is very likely that the fields of view of spatially proximal camera nodes would be highly correlated as the views among these nodes are overlapped. This causes the nodes to send redundant information, wasting network resources. This paper proposes a novel geometrical correlation model to estimate the correlation characteristics among randomly deployed heterogeneous camera nodes. The proposed model is mathematically analyzed and a closed-form analytical correlation function is provided. The results demonstrate that the proposed model is appropriate for applications with heterogeneous camera nodes being deployed in a random manner and it can also be efficiently used in applications that require a bounded distortion-based camera selection strategy. That is, selecting the minimum number of nodes that deliver the maximum amount of information with the least distortion.