Proposed New AV-Type Test-Bed for Accurate and Reliable Fish-Eye Lens Camera Self-Calibration.

Area of Science:

• Computer Vision
• Optical Engineering
• Metrology

Background:

• Fish-eye cameras offer wide fields of view but suffer from significant geometric distortion.
• Accurate interpretation of this distortion requires robust self-calibration procedures.
• Existing self-calibration methods can be limited by image point distribution and parameter correlations.

Purpose of the Study:

• To introduce a novel AV-type test-bed for fish-eye camera self-calibration.
• To evaluate the test-bed's ability to achieve balanced image point distribution and minimize parameter correlations.
• To validate the effectiveness of the AV-type test-bed using simulations and real-world data.

Main Methods:

• Development of the AV-type test-bed for controlled image acquisition.
• Implementation of self-calibration procedures using the AV-type test-bed with multiple projection models and datasets.
• Validation through simulation experiments and real-world data collection with different cameras.

Main Results:

• Simulation experiments demonstrated root mean square residuals (RMS-residuals) below 0.5 pixels across various conditions.
• Real-world experiments achieved high calibration accuracy, with minimum RMS-residuals of 0.39 pixels.
• The proposed test-bed consistently yielded lower RMS-residuals compared to standard methods.

Conclusions:

• The AV-type test-bed is highly effective for fish-eye camera self-calibration.
• The test-bed facilitates accurate geometric distortion correction.
• This approach enhances the reliability and precision of fish-eye camera systems.