Vision Mesh: A Novel Video Sensor Networks Platform for Water ConservancyEngineering

摘要

Video is an important medium for the observationof a variety of phenomena in the physical world such as forwater conservancy engineering. Video sensor networks (VSN)consists of a large number of sensor nodes with cheap CMOScameras, therefore, it has the ability of large scale visualmonitoring. However, a kind of powerful and scalable platformis urgent for simulation and development of VSN. This paperproposed a novel and scalable video sensor networks platformnamed as Vision Mesh for water conservancy engineering.Vision Mesh is composed of a mass of image or video sensornodes – Vision Motes, with which multi-view image or videoinformation of FOV (Field of View) can be acquired simply.Vision Mote is built with Atmel ARM9 CPU and operated onLinux operating system, with TI CC2430 Zigbee module takenas wireless transceiver. Furthermore, OpenCV machine visionlib is migrated to Vision Mesh platform so as to improve videoprocessing ability, therefore, Vision Mesh has the ability ofimage and video processing and is of strong scalability toextend performance. In this paper, we provide an overview ofVision Mesh architecture and provide an insightful platformfor VSN.