FIDUCIAL MARKER SYSTEMS OVERVIEW AND EMPIRICAL ANALYSIS OF ARUCO, APRILTAG AND CCTAG

The present paper advances an overview and comparison of existing fiducial marker systems. A fiducial marker is a 2D symbol that can be printed on a sheet of paper and glued to any flat surface to add distinctive characteristics to an environment. A system of fiducial markers consists of a set of un...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Scientific Bulletin. Series C, Electrical Engineering and Computer Science Electrical Engineering and Computer Science, 2023-01 (2), p.49
Hauptverfasser: Patru, George-Cristian, Pırvan, Alina-Irina, Rosner, Daniel, Rughinis, Razvan-Victor
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The present paper advances an overview and comparison of existing fiducial marker systems. A fiducial marker is a 2D symbol that can be printed on a sheet of paper and glued to any flat surface to add distinctive characteristics to an environment. A system of fiducial markers consists of a set of unique fiducial markers that contain information and an algorithm that can recognize and process the respective information using images and videos. This paper describes an evaluation of the ArUco, AprilTag, and CCTag marker systems. With the help of a minimalist hardware system, we obtained a data set consisting of 13200 images. Based on the data set, we evaluated the three marker systems according to the following criteria: the distance between the camera and the marker, the intensity of light, the movement speed of the camera, the resistance to occlusion, and the performance of the algorithms. The results obtained show that ArUco is the fastest, detecting the marker in 2 ms. Furthermore, the results illustrate that the CCTag is resistant to occlusion and performs very well in scenarios that simulate real-time detection, that is, scenarios with blurred pictures.
ISSN:2286-3540