Research On RGB-D-based Indoor Environment Building And Stereoscopic Display Technology

With the continuous development of science and technology,Virtual Reality(VR)technology has become a hot topic in the field of computer research.It has not only attracted the attention of scientific researchers,but also received a high degree of attention in the consumer field.Under this system,a virtual world can be created and felt.The most basic is the visual experience.Besides,the ideal virtual reality technology can provide users with a full range of sensory functions.Simultaneous Localization and Mapping(SLAM)is the core problem of robot navigation,one of the most popular topic in the near future that domestic and foreign scholars all work hard on the key issues.The V-SLAM based on the RGB-D depth vision sensor solves the problems of low efficiency and poor precision in monocular and binocular camera construction.With the emergence of consumer-grade RGB-D sensors,more developers join into the ranks of SLAM research.This topic integrates V-SLAM and virtual reality technology.With parts of the V-SLAM process are improved,an accurate and robust environmental reconstruction method is finally obtained.According to the results of a large-scale environmental reconstruction,a virtual reality scenario with immersion experience can be created.Firstly,the front-end of SLAM based on an RGB-D camera is further studied and designed.The depth camera structure and depth information measurement principle are studied and the camera's internal parameters are calibrated.Combining the iterative closest point and depth image information,a camera pose estimation method for fusion geometry information and RGB-D information is established,solved by Lie algebra.And the accuracy of the method is verified by the experiments.Secondly,study the construction principle of the TSDF map,and establish a model by integrating it with the SLAM method.For the limitation of “work space”,the camera motion area is expanded.According to the pose of the camera,the depth map is continuously integrated into the model.The results show that the multi-frame fusion model is significantly better than the single-frame model.Based on this,this paper proposed the addition of loop detection based on bag of words for reconstruction algorithms.This optimized the trajectory and modify the point cloud and model.Multiple constraints are set to reduce the possibility of loop errors,and experimentally prove that the method improves the accuracy of modeling.Then,combined with the results of 3D reconstruction of the scene,the key technologies of virtual reality are studied.The immersive environment perception is given to the user through the reproduction of the scene.Based on socket network communication,cross-system data transmission is completed,and Unity is used to perform secondary development of the VR all-in-one device.Screen split screen processing is performed based on binocular stereo vision to repair display distortion and realize display of a three-dimensional scene on a VR-integrated machine.Finally,using the Kinect sensor to conduct field experiments,the local and largescale scenes were not collected and the reliability and accuracy of the overall algorithm were verified.In addition,the TUM data set is used to evaluate the algorithm and the absolute error of the trajectory is analyzed to verify that the algorithm has good accuracy.