Probabilistic Camera-to-Kinematic Model Calibration for Long-Reach Robotic Manipulators in Unknown Environments

In this paper, we present a methodology for extrinsic calibration of a camera attached to a long-reach manipulator in an unknown environment. The methodology comprises coarse frame alignment and fine matching based on probabilistic point set registration. The coarse frame alignment is based on the k...

Full description

Saved in:
Bibliographic Details
Published in2022 IEEE 17th International Conference on Advanced Motion Control (AMC) pp. 48 - 55
Main Authors Makinen, Petri, Mustalahti, Pauli, Launis, Sirpa, Mattila, Jouni
Format Conference Proceeding
LanguageEnglish
Published IEEE 18.02.2022
Subjects
Cameras
iterative methods
Location awareness
Machine vision
Manipulators
Probabilistic logic
robot vision systems
Simultaneous localization and mapping
Visualization
Online AccessGet full text

Cover

More Information
Summary:In this paper, we present a methodology for extrinsic calibration of a camera attached to a long-reach manipulator in an unknown environment. The methodology comprises coarse frame alignment and fine matching based on probabilistic point set registration. The coarse frame alignment is based on the known initial pose and assists in the fine matching step, which is based on robust generalized point set registration that utilizes position and orientation data. Comparison with other methods utilizing only position data is provided. The first 6 DOF point set is obtained using a SLAM algorithm running on a camera attached near the tip of a manipulator, whereas the second point set is obtained using a kinematic model and joint encoders. Real-time experiments and a use case are presented. The results demonstrate that the proposed methodology is suited for the application, and that it can be useful in operations requiring precise visual measurements obtained near the tip of the manipulator.
ISSN:1943-6580
DOI:10.1109/AMC51637.2022.9729259