High-precision camera calibration via perspective distortion correction using a hybrid chessboard-circle target.

Area of Science:

• Computer Vision
• Optical Metrology

Background:

• Traditional camera calibration methods using chessboard or circular targets have limitations.
• Chessboard targets offer easy corner detection but low accuracy; circular targets have high accuracy but suffer from perspective distortion (eccentricity error).

Purpose of the Study:

• To develop a high-precision camera calibration method overcoming limitations of existing techniques.
• To introduce a novel hybrid calibration board combining chessboard and circular marker features.

Main Methods:

• A hybrid calibration board with circular markers on a chessboard pattern is proposed.
• Chessboard corners detect homography for image rectification, transforming elliptical projections to circles.
• Image moments extract corrected marker centers, and inverse perspective transformation refines pixel coordinates, mitigating eccentricity error.
• Zhang's method is applied with refined coordinates for camera calibration.

Main Results:

• The proposed method achieved an average reprojection error of 0.0365 pixels.
• This represents a 51.13% improvement over traditional chessboard patterns.
• A 35.05% improvement was observed compared to traditional circular patterns.

Conclusions:

• The novel hybrid calibration board and rectification method significantly enhance camera calibration precision.
• The approach effectively mitigates perspective-induced errors, leading to more accurate optical measurements.
• This method offers a superior alternative for precision optical measurement systems requiring high-accuracy camera calibration.