Deployment of visual sensor networks using a graph-based approach

We present a method for automatic deployment of visual sensor networks, based on a polygonal mesh model of the task. This method uses a graph-based approach to ensure an optimal level of visual overlap between the sensors. In this brief our method is further leveraged by increasing the size of the solution space, which in turn moves the solution closer to the optimal state. The increase in computational cost incurred by the extended solution space is handled by taking advantage of some occlusion properties of a convex partitioning of the task. Simulations are used to demonstrate the effectiveness of the approach.