Bad Photogrammetry: Using Nonoptimal or Archived Photographs for Constructing 3D Models
Archaeology frequently adopts new technology for fieldwork, but often this technology is of little use for prior fieldwork. One relatively new technology is using photogrammetry for constructing 3D models of features and artifacts. One of the many advantages of creating 3D models of archaeological features is the ability to merge models of multiple features to create a more complete view of an excavation. Often the excavation of sites occurs over multiple field seasons and occasionally different, adjacent features will not be open at the same time. Photogrammetry is limited by the quality of the photography, and there are several requirements for capturing optimal photographs. This paper explores the challenges and results of using nonoptimal photos from past excavations, including archived photographs, to create 3D models and produce useful results. We test the feasibility of attempting this process on several sites in the United States Southwest using relatively recent excavation photographs and with archived photographs from as far back as the 1930s. We focus on the workflow used to create the 3D models and discuss what situations failed to generate useful results and which were successful. We also discuss the benefits of creating 3D models from past excavations.
(Robert J. Bischoff, James R. Allison)