An Open-Source Photogrammetry Workflow for Reconstructing 3D Models

Area of Science:

• Biological Sciences
• Computational Biology
• Morphology

Background:

• Accurate and economical 3D biological models are crucial for organismal biology research.
• Structure-from-motion (SFM) photogrammetry offers a flexible and low-cost approach to 3D modeling.
• Reliance on expensive commercial software and insufficient accuracy assessments hinder photogrammetry adoption in biology.

Purpose of the Study:

• To present an affordable, open-source SFM photogrammetry pipeline using OpenDroneMap (ODM) and WebODM.
• To evaluate the geometric accuracy of ODM-derived 3D models by comparing them to microCT scans.
• To provide guidance for researchers on utilizing this pipeline for 3D biological model generation.

Main Methods:

• Developed and implemented an SFM photogrammetry pipeline utilizing the open-source OpenDroneMap (ODM) software.
• Acquired 3D models of 15 *Aplodontia rufa* (mountain beaver) skulls using the ODM pipeline.
• Assessed geometric accuracy by comparing ODM-derived models against microCT-scanned models of the same specimens.

Main Results:

• The ODM pipeline produced 3D models with sufficient geometric accuracy (errors ~2%) for gross metric and morphometric analyses.
• Consistent patterns of shape variation were captured in both ODM and microCT modalities.
• Subtle differences between modalities can affect landmark placement and downstream shape analysis, especially with low sample variance.

Conclusions:

• The ODM pipeline offers a cost-effective and accurate solution for 3D biological modeling for many applications.
• Researchers should not combine photogrammetric and microCT-derived models for geometric morphometric analysis due to potential inaccuracies.
• A pilot accuracy test is recommended before using SFM photogrammetry models for detailed morphometric studies.