Photomodeling vs. traditional 3D data capture of cultural heritage artifacts

Gabriele GUIDI¹ / Bernard FRISCHER²
(¹Politecnico di Milano, Italy / ²Indiana University, USA)

Keywords: 3D modeling, photomodeling

Abstract:
In recent years, photomodeling solutions for creating 3D models have matured to the point where they have become part of the Virtual Heritage toolkit. We will begin our paper by showing some recent examples of photomodeling of ancient sculpture from Hadrian’s Villa (Tivoli, Italy) and of the archaeological site of Cosa in Tuscany (Italy). Photomodeling has the advantage of requiring only a digital camera, not a dedicated piece of equipment such as a laser or structured-light scanner. This difference results in an immediate savings in cost since even a professional-level digital camera is less expensive than a scanner. The time needed for photographic data capture is typically measured in minutes, as compared to the hours needed by users of the dedicated devices. There is a price to be paid for this convenience: the 3D model made photographically does not have the same complexity as one made by laser or structured-light scanning. The former typically consist of 4 to 15 million polygons whereas the latter can measure in the hundreds of millions. Models in the range of 4 to 15 million polygons are useful for visualization, but they are usually not sufficiently detailed for offering the level of detail required for scientific documentation of cultural heritage artifacts. This paper reports on research into the question of how the model produced photographically —whatever its size and resolution—compares in terms of accuracy to a model produced by a dedicated device. Our test case will be the statue of Caligula in the Virginia Museum of Fine Arts, of which we have 3D models made in the traditional way (Faro Arm with data processed by Polyworks) and through photomodeling (Agisoft Photoscan and Autodesk Recap).