Multi-Camera 3D Digital Image Correlation with Pointwise-Optimized Model-Based Stereo Pairing

Area of Science:

• Engineering
• Measurement Science
• Computer Vision

Background:

• Dynamic deformation measurement (DDM) is crucial for infrastructure and industry.
• Multi-camera digital image correlation (MC-DIC) offers non-contact 3D DDM but suffers from inefficient camera pairing and over-grouping.
• Existing methods lead to complex setups and inaccurate deformation measurements.

Purpose of the Study:

• To introduce a novel MC-DIC method for precise and robust dynamic deformation measurement.
• To overcome the limitations of manual or simplistic camera pairing strategies in MC-DIC.
• To enable accurate DDM of semi-rigid objects even with complex backgrounds.

Main Methods:

• Proposed a pointwise-optimized model-based stereo pairing (MPMC-DIC) method.
• Developed an automated system for evaluating and selecting optimal camera pairs based on five key factors.
• Utilized 3D model information and calibrated camera data for pair selection.

Main Results:

• Achieved high-precision DDM with an accuracy of 0.03 mm for displacements up to 5.0 mm in a cylinder experiment.
• Demonstrated robustness against cluttered backgrounds using a 2x4 camera array.
• Successfully performed wide-range dynamic deformation analysis on a vibrating PC speaker with complex geometry.

Conclusions:

• MPMC-DIC effectively overcomes the over-grouping problem in MC-DIC.
• The proposed method enhances accuracy and robustness in dynamic deformation measurement.
• MPMC-DIC is suitable for analyzing dynamic deformations in complex scenarios and geometries.