Enhancing AprilTags Detection Range in ROS-Based Vision Systems Using Reflective Panels

Jiekai Wu, Ye Liu, Chung Shue Chen

Research output: Chapter in Book or Report/Conference proceedingConference Proceedingpeer-review

Abstract

This study reports a passive enhancement approach to increase the detection range and reliability of AprilTag fiducial markers in a ROS (Robot Operating System)-based vision system by placing simple reflective panels around the marker. We designed and fabricated two reflectors (two-wing and four-wing, constructed from foam board and household aluminum foil) that were adjustable over panel-to-tag angles between 0° and 90°. We evaluated their performance under various lighting conditions, ranging from no overhead lighting (0% ON) to having overhead lighting with full intensity (100% ON), using a Logitech C270 webcam and the standard ROS package apriltag-ros. From our experiment and observations, we see that the optimal panel inclination is 45° for the detection range and reliability. We investigate the performance of the four-wing and two-wing reflectors at 45° considering various lighting conditions. When compared with no reflector (the baseline scheme), the two-wing reflector and four-wing reflector can increase the detection range by 16.3% and 16.6% on average, respectively. Results show that the four-wing reflector is very effective under most lighting conditions (from 25% ON to 100% ON), achieving a mean detection distance of 6.89 m which is promising for many practical applications. On the other hand, when ambient light (0% ON) is the only lighting source, the two-wing reflector performs better in avoiding self-occlusion. We conclude that low-cost, easy-to-install reflective panels can serve as effective passive light amplifiers for AprilTags that enhance tag detectability in real-world scenarios without requiring active device, additional power or high system complexity.
Original languageEnglish
Title of host publicationThe 10th Optoelectronics Global Conference (OGC 2025)
Publication statusAccepted/In press - 24 Jul 2025

Fingerprint

Dive into the research topics of 'Enhancing AprilTags Detection Range in ROS-Based Vision Systems Using Reflective Panels'. Together they form a unique fingerprint.

Cite this