Spatial light path analysis and calibration of four-mirror-based monocular stereo vision

Area of Science:

• Computer Vision
• Optical Engineering
• Robotics

Background:

• Four-mirror-based monocular stereo vision (FMSV) systems require precise layout and calibration for accurate performance.
• Existing 2D light path analysis is insufficient for fully characterizing FMSV imaging parameters.
• Traditional calibration methods often impose constraints, such as optical axis perpendicularity to the calibration target.

Purpose of the Study:

• To develop a spatial light path analysis method for refining FMSV system design.
• To establish a depth of field (DoF)-distortion equal-partition model for improved calibration.
• To propose an accurate and practical FMSV calibration method that removes traditional constraints.

Main Methods:

• A spatial light path model was developed to overcome limitations of 2D analysis.
• A depth of field (DoF)-distortion equal-partition model was established, considering lens distortion.
• A novel FMSV calibration method was proposed, removing the perpendicularity constraint.
• A high-accuracy, high-portability FMSV system was constructed and calibrated using the developed techniques.

Main Results:

• The spatial light path analysis refined the FMSV system design.
• The proposed calibration method achieved an average error of 0.0298 mm in vision-reconstructed distance.
• Robot path accuracy evaluated by the FMSV system showed high accuracy of 0.031 mm when compared to laser-tracker measurements.

Conclusions:

• The proposed spatial light path analysis and calibration methods significantly enhance the accuracy and feasibility of FMSV systems.
• The developed FMSV system demonstrates high precision suitable for applications requiring accurate distance and path measurements.
• The novel calibration technique offers a more practical and less constrained approach for FMSV system implementation.