Font Size: a A A

A Study On The Technology Of Node Location In Video Sensor Networks

Posted on:2010-04-25Degree:MasterType:Thesis
Country:ChinaCandidate:K QiFull Text:PDF
GTID:2178330338475937Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Video sensor networks is a kind of distributed sensor network which are make up of computing, storage and communicating capabilities of video sensor nodes. Node localization in video sensor networks is not only the premise of target identification, monitoring, tracking, and many other applications, but also is the theoretical basis of achieving self-configuring network topology , improving routing efficiency and network coverage quality for network functions. Therefore, Node localization is one of the hot study of video sensor networks.The research background, related content and significance about video sensor networks and node positioning technology are introduced in this paper. Around the problem about nodes position in Visual sensor network, target location based on computer vision algorithms were also discussed in this paper. And then this paper implements a two-dimensional coordinates of the node positioning algorithm based on Camera Sensor. For the first the feasibility and the scene should be study, determine the physical parameters of Camera Sensor, use the of pinhole imaging model to Collect the image form unknown node by three or more anchor node image information in its sight, Secondly, analysis the image relationship between their image points and object points, at last achieve self-positioning problem and analysis node localization algorithm error based on two-dimensional coordinates of the Camera Sensor.In video sensor networks, Unknown node and anchor node may not be in a two-dimensional coordinate system, but in a three-dimensional space, this paper propose a positioning algorithm in a three-dimensional coordinate system based on the Camera Sensor node. The video sensor nodes in a three-dimensional space broken down into three two-dimensional plane, in each plane it can use two-dimensional coordinates of the node positioning algorithm to solve the three-dimensional localization problem of the Camera Sensor node.Finally, the hardware nodes of video-based sensor networks CMUCAM3 platform are realized and improved in this paper. The platform can capture resolution of 352×288 (RGB) digital image information, transfer the high-speed USB data. In real scenarios, CMUCAM3 node platform acquire their field scope of vision anchor node image information to complete the task of node location by Camera Sensor node localization algorithm.Experiments show that video sensor nodes CMUCAM3 platform can acquire the image data, processing and other functions. In 600×600cm experiment scenarios, CMUCAM3 nodes can complete self-positioning task based on the Camera Sensor node localization algorithm, the positioning accuracy is high and positioning error of up to cm-level.
Keywords/Search Tags:Video sensor networks, target positioning technology, physical parameters, positioning accuracy
PDF Full Text Request
Related items