On Scaling Distributed Low-Power Wireless Image Sensors

The power cost associated with communicating images across a wireless sensor networks is high. Therefore, we provide a method of organizing a large distributed system through analysis of its sensing capabilities to efficiently communicate features of images within the sensor field. Concurrent detections of targets in a sensor field are used to analyze the field of view of image sensors. Link weights are established to provide intelligent communication behavior. Using these link weights, sensor nodes communicate to those that need target features to estimate a 3D track. They also communicate to warn sensor nodes of an approaching target. We illustrate the feasibility of this approach through simulations and deployed experiments.