On-site semi-automatic calibration and registration of a projector-camera system using arbitrary objects with known...

Area of Science:

• Computer Vision
• Robotics
• Computer Graphics

Background:

• Projector-camera systems augment physical objects with virtual textures.
• Precise calibration is crucial but historically complex and time-consuming.
• Existing self-calibration methods lack global scale and are sensitive to data.

Purpose of the Study:

• To develop a semi-automatic calibration approach for projector-camera systems.
• To recover global scale, a limitation of current auto-calibration methods.
• To improve the accuracy of intrinsic and extrinsic calibration parameters.

Main Methods:

• Combines surface registration with bundle adjustment optimization.
• Utilizes points reconstructed from structured light projections.
• Refines solutions derived from fundamental matrix decomposition.

Main Results:

• Successfully estimates global scale robustly in simulations and real-world experiments.
• Improves incorrectly guessed intrinsic and extrinsic calibration parameters.
• Outperforms comparable metric rectification algorithms.

Conclusions:

• The proposed semi-automatic method offers a robust solution for projector-camera system calibration.
• It overcomes the global scale limitation of self-calibration techniques.
• This approach enhances accuracy and efficiency in augmented reality applications.