Distorted pinhole camera modeling and calibration.

Area of Science:

• Optics and Photonics
• Computer Vision
• Metrology

Background:

• Traditional camera calibration relies on the pinhole model, often extended for lens distortion.
• The interdependence between pinhole and distortion models in conventional methods leads to systematic calibration errors.
• Accurate camera modeling is crucial for various optical and metrology applications.

Purpose of the Study:

• To present a unifying distorted pinhole camera model that encompasses telecentric, distortion-free pinhole, and radial lens distortion as special cases.
• To develop and evaluate an iterative calibration method based on the proposed unified model.
• To address systematic errors in standard camera calibration techniques.

Main Methods:

• Derivation of a unifying distorted pinhole camera model.
• Development of an iterative calibration algorithm utilizing the new model.
• Experimental validation using a camera with significant radial distortion.
• Comparative analysis against standard and fisheye camera models using commercial software.

Main Results:

• The proposed distorted pinhole model accurately represents various camera configurations, including telecentric and distortion-free scenarios.
• The iterative calibration method based on the new model demonstrates superior performance over the standard pinhole model.
• Achieved calibration accuracy comparable to specialized fisheye models.
• Experimental results confirm the model's effectiveness in reducing systematic errors.

Conclusions:

• The unifying distorted pinhole camera model offers a versatile and accurate solution for camera calibration.
• The proposed iterative calibration method overcomes limitations of standard approaches, enhancing precision in optical metrology.
• This research provides a robust tool for diverse applications requiring precise camera characterization.