Using Nonoptimal or Archival Photographs for Constructing 3D Models

Abstract

Photogrammetry is well on its way to becoming a standard part of archaeology. As helpful as new technology can be, it is not always applicable to past excavations. We recommend following best practices for all current work, but photographs from prior excavations are generally not ideal for photogrammetry. We demonstrate the utility of photogrammetry under various, problematic conditions using test cases from three archaeological sites excavated between the 1930s and the current decade. We discuss some of the successes, challenges, and conditions where we succeeded or failed to create a viable 3D model. Our results indicate that successful and accurate 3D models can be created from excavation photographs taken under a variety of circumstances and without accompanying camera metadata. The major limiting factor we found was lack of overlapping coverage of the subject, which proved problematic to a greater or lesser degree in all of our case studies.