Improvements to Thin-Sheet 3D LiDAR Fiducial Tag Localization

The intensity image-based LiDAR fiducial marker system (IFM) proposes an algorithm to localize 3D fiducials of thin-sheet tags, using patterns compatible with AprilTag and ArUco, which can be attached to other surfaces without impacting the 3D environment. Unfortunately, due to the adoption of 3D-to-2D spherical projection, IFM exhibits two limitations: 1) IFM can only detect fiducials in a single-view point cloud and does not apply to a 3D LiDAR map; and 2) as the distance between the tag and the LiDAR increases, the projection size of the tag decreases until it is too small to be detected. In this paper, aiming to tackle the limitations and benefit downstream tasks such as 3D map merging, we develop an algorithm to improve the localization of thin-sheet 3D LiDAR fiducial tags. Given a 3D point cloud, which can serve as a 3D map, with intensity information, our method automatically outputs tag poses (labeled by ID number) and vertex locations (labeled by index) with respect to the global coordinate system. In particular, we design a new pipeline that gradually analyzes the 3D point cloud of the map from the intensity and geometry perspectives, extracting potential tag-containing point clusters. Then, we introduce an intermediate-plane-based method to further check if each potential cluster has a tag and compute the vertex locations and tag pose if found. We conduct both qualitative and quantitative experiments to demonstrate that the proposed method addresses the limitations of IFM while achieving better accuracy. The open-source implementation of this work is available at: https://github.com/York-SDCNLab/Marker-Detection-General.