A Stable, Efficient, and High-Precision Non-Coplanar Calibration Method: Applied for Multi-Camera-Based Stereo Vision...

Area of Science:

• Computer Vision
• Metrology
• Optical Engineering

Background:

• Traditional non-coplanar calibration methods struggle with accuracy and usability in multi-camera stereo vision.
• Existing techniques like Tsai's method present limitations for precise 3D measurements.

Purpose of the Study:

• To develop novel non-coplanar calibration methods for monocular and binocular camera systems.
• To enhance the accuracy and stability of stereo vision calibration through improved feature point extraction.
• To enable simultaneous calibration of intrinsic and extrinsic parameters in multi-camera systems.

Main Methods:

• Established a novel mathematical model based on projective and matrix transformation theory.
• Developed a new non-coplanar calibration method for monocular and binocular systems.
• Proposed a circular feature point extraction method using region Otsu algorithm and radial section scanning.

Main Results:

• The novel calibration methods demonstrate superior accuracy and ease of operation compared to Tsai's and Zhang's methods.
• The feature point extraction algorithm offers high precision and stability, improving overall calibration accuracy.
• Simultaneous calibration of intrinsic and extrinsic parameters for multi-camera systems is achieved.

Conclusions:

• The proposed non-coplanar calibration and feature extraction methods provide high accuracy and stability for stereo vision measurements.
• These advancements are suitable for complex shape and deformation measurements, including stereo-digital image correlation (stereo-DIC).