Detection and tracking of region-based evolving targets in sensor networks

Abstract

Sensor networks have emerged as a promising technology for large-scale, real-time monitoring in complex environments. This paper develops in-network algorithms for the detection and tracking of region-based targets and events. We deal with such issues as network topology, event boundary detection, region merging, and region simplification and smoothing. Simulation experiments confirmed that our algorithms are efficient and accurate.