| Title |
Robust and Fast Collaborative Augmented Reality Framework Based on Monocular SLAM |
| Authors |
(Jae-Won Kam) ; (Han-Sol Kim) ; (Sang-Jun Lee) ; (Sung-Soo Hwang) |
| DOI |
https://doi.org/10.5573/IEIESPC.2020.9.4.325 |
| Keywords |
Collaborative AR; Monocular SLAM; Map fusion; Map evaluation |
| Abstract |
When team members need to share the same virtual object in the same workspace, it is difficult to set the proper viewpoint of the virtual object for all the team members, which is called the pose estimation problem. In this paper, we propose a collaborative augmented reality (AR) framework based on monocular simultaneous localization and mapping (SLAM) to address these issues. The proposed framework consists of a map fusion module and a map quality evaluation module. The map fusion module solves the problem of relative pose estimation on collaborative AR by performing fast overlap detection and optimization. Thanks to this module, each team member can generate a map independently, and the maps can be quickly aligned and merged into a single global map. The map quality evaluation module assesses the quality of the map and provides feedback to team members. It helps each team member create a robust map for pose estimation. In the experiments, we show how the map fusion module can fuse maps with high speed and accuracy. In addition, we show that maps validated with the map quality evaluation module are robust to pose estimation. We also demonstrate the feasibility of the framework by implementing a simple AR application through the proposed framework. |