Parameter-free radial distortion correction with center of distortion estimation

Area of Science:

• Computer Vision
• Computational Imaging
• Optical Engineering

Background:

• Accurate camera calibration is crucial for various computer vision tasks.
• Existing methods often rely on specific radial distortion models or iterative processes.
• Identifying the precise center of radial distortion is often overlooked but critical.

Purpose of the Study:

• To develop a novel, simultaneous calibration method for radial distortion and internal camera parameters.
• To create a noniterative, parameter-free approach for broad camera applicability.
• To accurately determine the center of radial distortion, independent of image or principal point.

Main Methods:

• Simultaneous calibration of radial distortion and internal camera parameters using a calibration grid in multiple images.
• Extension of Zhang's popular calibration method to incorporate radial distortion determination.
• Noniterative, parameter-free radial distortion modeling.

Main Results:

• The proposed method is rapid and immune to local minima due to its noniterative nature.
• It accurately determines radial distortion without assuming a specific model, applicable to diverse lenses (narrow-angle to fish-eye).
• The computed center of radial distortion can be significantly displaced from the image center or principal point, impacting calibration accuracy.

Conclusions:

• The method offers an efficient and robust solution for camera radial distortion and internal parameter calibration.
• Its parameter-free nature and accurate center determination enhance applicability and performance across various camera systems.
• This technique provides a valuable tool for improving the accuracy of computer vision applications reliant on precise camera models.