The Design And Implementation Of Large-scaleUser TrackingSysten Based On Multi Kinects

With the rapid development of virtual reality technology,tracking positioning and motion recognition for users are important technologies for virtual reality.Various devices for positioning and motion capture have also emerged.Vive developed by HTC and Kinect developed by Microsoft and other devices can perform high precision positioning and motion recognition for users.These devices have a good user experience and can well support virtual reality technology,but because of the limitations of the devices,these devices can only be used in a small range of scenarios,these devices will be powerless in a wide range of virtual reality scenarios,which limits the device usage scenarios,affecting the user's immersion.Therefore,how to use the existing equipment and technology to break through the limitations of equipment to provide better technical services for virtual reality is very meaningful.Based on the above analysis,this paper designed and implemented a multi-user tracking system based on multi-Kinect to track and locate multiple users in a real scene,visualize the user's location information,and provide services for the virtual reality scene by providing user location information.This paper analyzes multi-camera calibration techniques and traditional single-camera tracking learning to achieve the calibration of Kinect,and-combines the technical features of K inect to expand to multiple Kinects so that the scope of capture space can be expanded or reduced as needed.The user's bone data space tracked by a single Kinect is approximately 12 square meters.Due to experimental environment limitations,this paper uses 4 Kinect as an example to expand the tracking range to approximately 36 square meters of space.Multple Kinect from different locations and perspectives obtaining users' information to solve the limits of the range of view caused by single Kinect,occlusion may occur between users have been processed,and how to processing the overlapping area of Kinects will be discussed,as well as how to solve multiple Kinect relay tracking,and other issues for processing.The multi-Kinect network module which is expansibility is implemented according to the above functions.The system provides a good visual interface for tracking results,defines and provides a good data interface,and provides services for virtual reality scenarios through the data interface.The system uses C/S architecture,each Kinect program preprocesses user information and sends data to the server side,then,the server side analyzes and integrates the received data of a plurality of Kinects,finally visualizes the user data and returns it to the virtual application scene.