Configuring video sensor network using a novel vision based technique

Abstract

Nowadays there is great enthusiasm in using video sensor networks in both civilian and military fields. This marks the location awareness of the node as a critical issue. In this paper we present a novel vision based technique that builds up a local relative coordinate system in one-plane video network without using GPS. One sensor node will present the core of the system and it will start locating some other 2 reference nodes that are randomly deployed using vision characteristic. They, in turn, will participate in locating the entire network with the help of multi-lateration method and the same vision characteristic. Some vision experiments are presented to demonstrate the feasibility of this method.