Large field-of-view calibration via circular targets and scale constraints

Area of Science:

• Computer Vision
• Metrology
• Geometric Calibration

Background:

• Traditional calibration of large field-of-view (FOV) vision systems is hindered by costly, difficult-to-manage precision targets.
• Manufacturing costs, operational challenges, and structural deformation limit current calibration methods.

Purpose of the Study:

• To develop a flexible and high-precision calibration method for large FOV vision systems.
• To overcome the limitations of traditional calibration techniques using large-scale precision targets.

Main Methods:

• Proposed a flexible calibration method integrating structure from motion (SfM) with geometric scale constraints.
• Utilized a moving camera to capture stationary circular control points, creating a multi-view geometric framework.
• Developed a global bundle adjustment optimization model incorporating the physical radius of circular targets as a constraint to resolve scale ambiguity in monocular SfM.

Main Results:

• Achieved high-precision estimation of intrinsic and extrinsic camera parameters.
• Demonstrated operational flexibility and robustness in calibration.
• Provided an efficient solution for large-scale industrial vision metrology.

Conclusions:

• The proposed SfM-based method offers a flexible, precise, and efficient alternative for calibrating large FOV vision systems.
• Eliminates the need for cumbersome large calibration boards, simplifying industrial applications.
• Successfully addresses scale ambiguity and projection eccentricity errors in monocular SfM.