Download: PDF.
“Underwater AprilTag SLAM and Calibration for High Precision Robot Localization” by E. Westman and M. Kaess, Robotics Institute. Carnegie Mellon University technical report CMU-RI-TR-18-43, Oct. 2018.
In this work we present a SLAM framework using the popular AprilTag fiducials for obtaining precise, drift-free pose estimates of an underwater vehicle. The framework also allows for simultaneous calibration of extrinsics between the camera and the vehicle odometry coordinate frame. The pose estimates may be used for various underwater tasks such as mapping and inspection, or as a ground-truth trajectory to evaluate the accuracy of other localization methods. We evaluate the effectiveness of the system with real-world experiments in a test-tank environment and demonstrate that it corrects drift that accumulates with dead-reckoning localization.
Download: PDF.
BibTeX entry:
@techreport{Westman18tr,
author = {E. Westman and M. Kaess},
title = {Underwater {A}pril{T}ag {SLAM} and Calibration for High
Precision Robot Localization},
institution = {Robotics Institute, Carnegie Mellon University},
number = {CMU-RI-TR-18-43},
month = oct,
year = {2018}
}