Group geometric calibration and rectification for circular multi-camera imaging system.

Area of Science:

• Computer Vision
• Geometric Calibration
• Multi-Camera Systems

Background:

• Circular multi-camera systems face calibration challenges due to absent shared feature points between opposite viewpoints.
• Existing calibration methods struggle with the unique geometry of circular arrangements.

Purpose of the Study:

• To propose a robust group geometric calibration and ideal rectification scheme for circular multi-camera imaging systems.
• To address the lack of preliminary shared feature points in opposite viewpoints.
• To extend the calibration method for both static and dynamic scenarios.

Main Methods:

• Designed a novel 3D calibration object with unique patterns to extract extrinsic parameters.
• Estimated the ideal camera projection matrix based on extracted extrinsic parameters.
• Extended the static calibration scheme to dynamic scenarios.

Main Results:

• The proposed method demonstrates superior performance compared to state-of-the-art calibration techniques.
• Experimental verification through 3D modeling applications shows significant improvements.
• The calibration scheme is effective for circular multi-camera imaging systems.

Conclusions:

• The developed group geometric calibration method provides an effective solution for circular multi-camera systems.
• The scheme's applicability is verified for 3D modeling and dynamic scenarios.
• This method is suitable for future interactive visual applications.