Defocused camera calibration with a conventional periodic target based on Fourier transform

Area of Science:

• Computer Vision
• Image Processing
• Metrology

Background:

• Accurate camera calibration is crucial for many computer vision applications.
• Traditional methods rely on sharp target images, limiting their use with defocused cameras.
• Defocus blur degrades image quality, hindering precise feature point extraction.

Purpose of the Study:

• To develop a novel camera calibration method robust to defocus blur.
• To enable accurate calibration using conventional periodic targets with defocused cameras.
• To extract features from the phase domain, overcoming intensity-based limitations.

Main Methods:

• Utilized a conventional periodic target for calibration.
• Applied Fourier transform to recover two crossed phase maps from the target image.
• Extracted feature points from the phase domain, rather than image intensity.
• Validated the method through simulations and experimental tests.

Main Results:

• Successfully recovered phase maps from periodic target images.
• Demonstrated feature point extraction robust to camera defocusing.
• Achieved accurate camera calibration using both focused and defocused images.
• Confirmed the efficacy of the phase-domain approach.

Conclusions:

• The proposed method offers accurate camera calibration for defocused systems.
• Phase-domain feature extraction provides robustness against defocus blur.
• This technique expands the applicability of camera calibration in challenging imaging conditions.