Calibration of multimodal 3D structured-light systems using digital features.

Area of Science:

• Computer Vision
• Metrology
• Optical Engineering

Background:

• Multimodal 3D imaging systems require precise calibration.
• Traditional calibration relies on physical targets, which are expensive and prone to defects.
• Challenges include feature detection across different modalities (e.g., reflections, limited resolution).

Purpose of the Study:

• To develop a cost-effective and robust calibration method for multimodal 3D imaging systems.
• To eliminate the need for specialized physical targets in calibration.
• To improve the accuracy and simplify the process of multimodal calibration.

Main Methods:

• Utilized a calibration setup with a screen, mirror, and auxiliary camera to generate digital features.
• Recovered intrinsic parameters of the visible camera.
• Determined intrinsic and extrinsic parameters for the projector and the additional-modality camera (demonstrated with a thermal camera).

Main Results:

• Achieved a 0.07 mm root-mean-square error in 3D reconstructions.
• Demonstrated comparable accuracy to conventional calibration techniques.
• Successfully eliminated the need for physical targets with specific features.

Conclusions:

• The proposed digital feature-based calibration method is effective for multimodal 3D imaging.
• This approach reduces calibration costs and avoids fabrication-related inaccuracies.
• Simplifies feature detection and enhances the reliability of multimodal 3D reconstruction.