Font Size: a A A

The Design And Implementation Of The Motion Capture And Comparison System Based On Kinects

Posted on:2018-12-02Degree:MasterType:Thesis
Country:ChinaCandidate:G Q MuFull Text:PDF
GTID:2348330515487080Subject:Computer technology
Abstract/Summary:PDF Full Text Request
Motion capture has important application values in a wide range of fields,such as somatic game,animation,education and training and other fields.At present,with the rapid development of the virtual reality(VR)technology,real-time motion capture has become an integral part of many VR systems to support the user to naturally interact with VR contents.Among different real-time motion capture methods,Kinect,which is pushed out by the Microsoft in 2010,are widely used because of its relatively low cost,good portability,operation reliability and other advantages.But the current capture range of Kinect has a certain limit,and only when the user faces to the Kinect can the motions be captured well.When the user sideways toward Kinect,it is hard for it to accurately collect the user's bone joint data,which limits its scope of application.The application background of this study is the construction of virtual TaiChi learning environment.We need to accurately capture and reproduce the learner's motion,and then compare it with the standard motion.For this purpose,this paper designed and developed a multi-Kinect-based 360-degree omni-directional motion capture system.After investigating and analysis of the existing motion capture system,especially the motion capture systems based on Optics,this paper developed a motion capture and analysis system based on Kinect network.The system used the skeleton data relay algorithm between Kinect and Kinect data filtering algorithm,taking advantage of the attribution of Kinect and traditional optical camera correction algorithm.Furthermore,the motion analysis part was complemented to compare the two sets of motion data,and feedback improvement advices for the nonconforming joint as well as a quantified standard score for the learning.The main part of the system was rendered by Unity3D.The server of Kinect data was developed as a WPF project when the calibration part was completed by MatLab.Finally,the different function parts were connected by net communication as a sheer system.The speed of development and the ability to maintain can be improved by applying different techs for different needs.
Keywords/Search Tags:motion capture, compare and analysis, Kinect, calibration, filter
PDF Full Text Request
Related items