Research On VR Interactive Input System Based On Vision And Inertial Navigation System

Since the advent of VR,its application has been deepening.With the impact of the COVID-19 pandemic in 2020 and the further popularization of 5G communications,emerging models such as VR+ training,AR real-time remote communication,unmanned driving and virtual live broadcasting are developing rapidly.People pay more and more attention to VR interactive technology with higher operation efficiency and better experience effect.Among them,high-precision VR interactive input system has important research significance for VR interactive technology.According to the characteristics of VR interactive input equipment,this paper proposes a pose detection method of input system based on binocular vision and INS data fusion,which improves the accuracy of position and pose measurement of VR interactive input equipment.The main contents of this paper are as follows.Firstly,the attitude angle and attitude matrix of INS system are analyzed,the INS position and attitude calculation process and its correction method are studied,the VR position and attitude measurement model based on INS is established,and the position and attitude calculation method of VR interactive input system based on Kalman filter with kinematic constraints is proposed.The performance of the proposed method is verified by attitude detection experiments.Secondly,the coordinate system of binocular camera position detection system is established,the calculation methods of internal and external parameters of the camera are analyzed,and the parallax measurement method of binocular camera is studied;The position measurement method based on binocular stereo vision input system is studied,the camera calibration and image correction methods are studied,and the feature point matching and three-dimensional position detection method of VR input system based on improved surf are proposed.The performance of the proposed method is verified by experiments.Thirdly,the detection system coordinate system is established,the system initialization method is analyzed,the pose measurement method of VR input system based on ins and binocular vision fusion is proposed,the binocular camera position solution method is studied,the ins time domain alignment method is studied,and the synchronization of camera and INS data in time domain is realized,A position detection fusion method of VR input system based on adaptive variable weighted kalman filter is proposed.The performance of the proposed method is verified by pose detection experiments.Finally,a pose detection prototype of VR input system based on ins and binocular vision is developed,the hardware circuit and software program of the system are designed,and the VR interactive input position and pose detection test method is designed.Through the position detection comparison test and pose detection comparison test,It is verified that the position and attitude detection method of VR interactive input system can effectively improve the position and attitude measurement accuracy.The research of this paper provides technical guidance for the accurate input of VR interaction,and has certain theoretical and practical value.It has important practical significance for the popularization of VR technology simulation and engineering application.