Cut from the same cloth : towards an automated trimming pipeline for archaeological 3d models

Abstract

By capturing excavated features with high precision and detail, photogrammetric 3D models provide an invaluable path toward reconstructing the inherently deconstructive process of archaeology, especially in a single-context methodology where each stratigraphic unit is documented with both topographic and photogrammetric techniques. Despite this potential, the practical use of photogrammetric models remains limited by the labor required to prepare them for deployment. Models must often be trimmed, decimated, organized, and exported before they can run in interpretive 3D visual environments, an enormous hurdle for underfunded projects hoping to integrate photogrammetry during site survey. This thesis attempts to bridge that gap by developing an open-source toolkit for trimming and decimating archaeological photogrammetry models, the two key processing steps required for single-context use. Relying on sample data from the Gabii Project, a Roman-era site outside of Rome and an early adopter of single-context photogrammetric recording, this project demonstrates how existing workflows might be adapted into an automated pipeline. While not wholly automatic, the tools lay the foundation for future edge-detection approaches using hand-corrected archaeological polygons as training data. The pipeline application and associated scripts will be made available in a GitHub repository for anyone attempting to integrate photogrammetric documentation into everyday field recording and interpretation.