Directionally Decomposing Structured Light for Projector Calibration.

Area of Science:

• Computer Vision
• Optical Engineering
• Robotics

Background:

• Accurate intrinsic projector calibration is crucial for projection mapping (PM), particularly in dynamic scenarios.
• Shallow depth-of-field (DOF) in projectors poses a significant challenge for precise calibration.
• Existing methods often struggle with projectors having limited DOF or when working distances are constrained.

Purpose of the Study:

• To develop a practical method for estimating projector intrinsic parameters that circumvents the limitations of shallow DOF.
• To present a calibration device enabling accurate calibration irrespective of projector focusing distance and aperture size.
• To validate the technique's efficacy for dynamic projection mapping applications.

Main Methods:

• A calibration device employing a flat-bed scanner and pinhole-array masks was designed.
• Projectors cast structured light patterns onto the device, with pinholes filtering specific rays.
• Extracted rays passing through pinholes were used to calibrate the projector using a standard pinhole camera model.

Main Results:

• The proposed technique successfully calibrated projectors with varying focusing distances and aperture sizes.
• Calibration accuracy was comparable to conventional methods.
• The method proved effective even for projectors positioned far from the target, where traditional methods fail.

Conclusions:

• The developed calibration device and method effectively address the DOF limitations in projector calibration.
• The technique provides intrinsic parameters suitable for high-accuracy dynamic projection mapping.
• This approach expands the applicability of projection mapping in scenarios with challenging projector configurations.