Calibration for Camera-Projector Pairs Using Spheres.

Area of Science:

• Computer Vision
• Robotics
• Geometric Calibration

Background:

• Traditional camera calibration methods often do not extend directly to projector calibration.
• Existing techniques for camera-projector systems can be complex and may not handle multiple sensor configurations effectively.
• Accurate calibration is crucial for 3D reconstruction and augmented reality applications.

Purpose of the Study:

• To develop a novel calibration algorithm for camera-projector systems utilizing spheres.
• To extend sphere-based calibration from cameras to projectors, enabling unified system calibration.
• To achieve robust calibration for single and multiple camera-projector pairs, particularly in surrounding sensor setups.

Main Methods:

• The algorithm treats projectors as inverse cameras, leveraging pole-polar relationships to derive sphere image conics on projector planes.
• It requires at least three image conics per device to compute intrinsic parameters.
• Extrinsic parameters are determined using sphere center positions across device coordinate frames.

Main Results:

• The proposed method successfully calibrates individual cameras and projectors.
• It demonstrates effective calibration for multiple camera-projector systems.
• Experimental results on synthetic and real data validate the algorithm's performance and accuracy.

Conclusions:

• The developed algorithm provides a unified and robust approach for calibrating camera-projector systems.
• It simplifies the calibration process for complex multi-sensor setups.
• This work contributes to improved 3D sensing and scene understanding in various applications.