Accuracy of fish-eye lens models.

Area of Science:

• Computer Vision
• Optical Engineering
• Geometric Optics

Background:

• Pinhole camera model is standard in computer vision.
• Real-world optical systems, especially fish-eye lenses, exhibit radial distortion.
• Existing models attempt to describe fish-eye lens displacement.

Purpose of the Study:

• To evaluate the accuracy of various fish-eye lens models.
• To present a novel method for extracting fish-eye camera lens curves.
• To compare empirical data with theoretical models.

Main Methods:

• Developed a method to extract the lens curve using perspective principles.
• Empirically derived the lens curve of a fish-eye camera.
• Compared several established fish-eye lens models against the derived curve.

Main Results:

• An empirically derived lens curve was successfully extracted.
• The accuracy of selected literature models was evaluated against this curve.
• Quantified the deviation of theoretical models from actual fish-eye lens behavior.

Conclusions:

• The study provides an empirical benchmark for fish-eye lens model validation.
• Highlights the importance of accurate lens distortion modeling in computer vision.
• Offers a method for characterizing real-world optical distortions.