Research And Design Of Some Key Technologies For Indoor VR Roaming System

The demand for virtual reality indoor scene applications is increasing,and people are increasingly demanding on issues that affect user experience,such as insufficient scene reality,limited screen fluency,and fatigue.Among the rendering technologies used in virtual reality,the traditional PBR technology becomes increasingly weak and faces a performance bottleneck in the face of increasingly sophisticated and increasingly complex scene vertex data.As a promising IBR technology,light field technology has begun to emerge in the field of virtual reality with the continuous follow-up of hardware in recent years.This technology has great advantages such as realistic rendering results,irrelevant computing overhead,and scene complexity,and conforms to human visual habits.However,the technology still has problems such as high acquisition costs,large data operations,and limited real-time performance.In response to the above problems,the main contributions of this paper are as follows:(1)A low-cost light field acquisition scheme suitable for indoor complex static scenes is propose.Only a traditional camera combined with a tripod can be used for acquisition,which avoids the problem of expensive and huge camera arrays.At the same time,it also inherits the advantages of high spatial resolution of the camera array solution.(2)A calibration and expansion method of data based on camera calibration and optical flow simulation is proposed as an optimization and supplement of low-cost acquisition schemes.Aiming at the problem that the accuracy of the individual shooting position in the solution is limited,the calibration data is used to correct the collected images by camera calibration;for the problem of limited acquisition density in the solution,analysis is performed from the perspective of image motion analysis,and the optical flow method is used Image synthesis realizes the expansion of the collected data set.(3)A light field reconstruction calculation optimization method based on rendering contributions is proposed to improve the real-time performance of light field reconstruction.By analyzing the light field reconstruction process and the resampling contribution for the final result,the calculation method and termination conditions of the resampling contribution is given to reduce the amount of reconstruction calculation.(4)Finally,based on the research results above,we developed a system for scene data pre-processing system and the indoor VR roaming system.This system is based on Visual Studio 2017 and Qt5 for system architecture design,algorithm modular encapsulation,and user interaction interface design.