Automatic Calibration of an Industrial RGB-D Camera Network Using Retroreflective Fiducial Markers

Area of Science:

• Robotics
• Computer Vision
• Sensor Networks

Background:

• Accurate sensor network calibration is crucial for 3D reconstruction and robotic applications.
• Existing methods often lack robustness or require invasive procedures.

Purpose of the Study:

• To develop an automatic, non-invasive, and robust calibration method for scalable RGB-D sensor networks.
• To achieve high accuracy in large-scale industrial environments.

Main Methods:

• Utilized retroreflective ArUco markers for precise marker detection.
• Employed the iterative closest point (ICP) algorithm for pose estimation.
• Implemented techniques to mitigate marker ambiguity and refine depth measurements.

Main Results:

• Achieved an average Euclidean error of 3 cm over distances up to 9.45 m.
• Demonstrated successful calibration of a six-sensor network in a 10x10x4 m industrial cell.
• The system proved robust against long-range and low-resolution detection issues.

Conclusions:

• The developed method provides an accurate and robust solution for RGB-D sensor network calibration.
• Open-source software and datasets are available, facilitating further research and application.