Calibration method of multi-projector display system with extra-large FOV and quantitative registration accuracy...

Area of Science:

• Computer Vision
• Display Technology
• Image Processing

Background:

• Multi-projector display systems require precise calibration for seamless visual splicing.
• Achieving accurate registration in extra-large field-of-view (FOV) displays presents significant challenges.

Purpose of the Study:

• To present a novel calibration method for seamless splicing in multi-projector display systems with extra-large FOV.
• To establish a quantitative registration accuracy analysis for such systems.

Main Methods:

• Utilized a wide-angle camera on a rotating optical system to scan the 360°x210° display.
• Established global longitude and latitude tables using a North Pole reference and inverse projection.
• Applied a three-frequency temporal unwrapping method for quantitative registration accuracy analysis.

Main Results:

• Developed a method to map each projector pixel to unique longitude and latitude values.
• Achieved a quantitative registration accuracy analysis for the multi-projector system.
• Experimental results demonstrated a registration error of less than 0.4 pixels.

Conclusions:

• The proposed calibration method effectively realizes seamless splicing for large FOV multi-projector displays.
• The quantitative registration analysis provides a reliable measure of system accuracy.
• The method offers a robust solution for achieving high-fidelity visual splicing in complex display environments.