In this project, we show how beacon-based indoor localization and additional environmental fingerprints like magnetic field data can be used to both accelerate and increase therobustness of Augmented Reality (AR) relocalization. We show how the combination of Visual Inertial Odometry (VIO) and beacons can be used to construct a dense indoor magnetic field map that can act as a fine-grained calibration for compasses to quickly determine a mobile device’s orientation. Unique to our approach is that we leverage accurate VIO trajectories to provide full vector magnetic field mapping that can be collectedand used with devices placed in any orientation.
