Benchmarking Localization for Augmented Reality in Large Scale Environments

Download PDF
Open Webpage

Shazia Gul, Wei Hong Lo, Steven Mills, Stefanie Zollmann

Published: 2023, Last Modified: 26 Feb 2026IVCNZ 2023EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Localization, which determines a user’s position and orientation, is crucial in accurately overlaying virtual content in Augmented Reality (AR). Our research investigates localization methods suitable for AR in large-scale environments and integrates them into a mobile system for evaluation. Specifically, we evaluate the performance of two state-of-the-art approaches, Expert Sample Consensus Applied to Camera Re-Localization (ESAC) and Accelerated Coordinate Encoding (ACE), focusing on their feasibility for AR in large-scale environments. As an experimental platform, we developed a client-server framework coupled with a mobile AR application for onsite feasibility testing. Using three large sample datasets — a stadium, a clocktower, and a courtyard; our approach processes a query image from the mobile AR client and provides pose computation using either ESAC or ACE. The pose data is then used in the mobile AR client to render 3D registered content.We conducted onsite feasibility testing with both methods, with ESAC evaluated using one and ten expert networks. We compare the accuracy and robustness of ESAC and ACE using the reprojection error as well as measuring computation times. Our results indicate that ESAC outperforms ACE for these types of datasets. In particular, we found that ten experts have higher success rates (83.8%, 95%, and 90%). When using one expert ESAC, the performance decreased (81%, 80%, and 85%) but still showed acceptable results around 80%. However, the ACE method shows reduced success rates (77%, 60%, and 75%) compared to ESAC. We also found that ESAC shows a smaller reprojection error compared to ACE. Our findings indicate that ESAC is promising and is suitable for localization for AR in a large-scale environment. This evaluation demonstrates the feasibility of the ESAC method in large-scale camera localization, providing insights for enhancing AR experiences.