Accurate Boundary Detection and Refinement for Continuous Objects in IoT Sensing Networks

Abstract

This article proposes a two-stage boundary detection and refinement mechanism in Internet of Things sensing networks, where boundary faces are generated by adopting planarization algorithms based on sensory data provided by backbone nodes. Candidate edge nodes, which may be observed more appropriate to serve as inner or outer boundary nodes, are discovered leveraging their sensory data estimated through adopting spatial interpolation methods, and their sensory data are gathered by mobile nodes with certain time constraints. This boundary refinement procedure is conducted in an iteration fashion, until no candidate edge nodes exist in, and thus, the refinement can hardly be further conducted upon original boundary faces. Experimental evaluation demonstrates that this technique can generate a more accurate boundary region for continuous objects than the state of the art.