Developing a Meta-AR Space Construction System and Mapping Coverage Visualization

Abstract

Visualizing the local environment for remote people is necessary for augmented reality (AR) and remote people between local and remote participants. Three dimensional (3D) reconstruction is a promising method for capturing the shapes and appearances in a local space, but accurate reconstruction requires high-precision sensors and high-performance computers. However, general-purpose smartphones and tablets have distance sensors like LiDAR, and we realized a meta-AR space based on such devices. Through synchronized AR and virtual spaces, local and remote users can collaborate in our meta-AR space. The performance and coverage of a smartphone-based 3D reconstruction system are two issues, though. First, 3D reconstruction using a conventional ray cast is slow; however, using a depth map, improvement is possible. Second, users are unable to see the captured space’s coverage; however, doing so is useful. The system cannot adequately reconstruct the space when the coverage is too low. This paper describes a quick 3D reconstruction system that supports the reconstruction with a depth map and mapping coverage visualization. Our experimental findings show its effectiveness, mapping coverage, and utility. The mapping coverage is adequate and our proposed method is 258 times faster than the ray cast method. Moreover, our mapping-coverage visualization increased the texture by 15%.