Object Localization with Multiplanar Fiducial Markers: Accurate Pose Estimation

Accurate object tracking is an important task in applications such as augmented/virtual reality and tracking of surgical instruments. This paper presents a novel approach for object pose estimation using a dodecahedron with pentagonal fiducial markers. Unlike traditional ArUco markers, which are squared, we propose a pentagonal marker that fits better in the dodecahedron figure. Our proposal improves marker detectability and enhances pose estimation with a novel pose refinement algorithm. Our experiments show the system's performance and the refinement algorithm's efficacy under different configurations.