Methodology of Object Reconstruction by Photogrammetry and Structured-Light Scanning for Industrial 3D Visualisation

Area of Science:

• Engineering
• Computer Science
• Metrology

Background:

• Digital twins require accurate 3D object reconstruction.
• Optical 3D scanning technologies are crucial for digitalization.
• Existing evaluation methods lack standardization.

Purpose of the Study:

• To introduce a unified methodology for evaluating optical 3D scanning technologies.
• To compare photogrammetry and structured-light scanning for 3D reconstruction.
• To ensure reproducible and comparable results for digital twin applications.

Main Methods:

• Developed a framework integrating data acquisition to quantitative validation.
• Implemented photogrammetry and structured-light scanning on the MiR100 robot.
• Utilized a reference CAD model for ground-truth comparison.
• Performed visual inspection, dimensional measurements, and statistical error analysis using MeshLab, CloudCompare, and MATLAB.

Main Results:

• Photogrammetry yielded high textural detail but suffered from geometric noise and scale drift (relative error > 10%).
• Structured-light scanning provided more stable and metrically accurate results.
• Structured-light scanning in scanner mode achieved the highest precision (mean deviation 17.4 mm, RMSE 26.8 mm, relative error 7.6%).

Conclusions:

• The proposed methodology provides a reproducible basis for 3D reconstruction accuracy evaluation.
• Structured-light scanning is more suitable for metrically accurate 3D object reconstruction than photogrammetry.
• This framework supports the integration of optimized digital models into digital twin environments.