Because one of our main goals is the scientifically-accurate documentation of ancient artifacts, the project used the newly designed, high-resolution STEM3D scanner. This triangulation-based, active laser scanner was designed by local Pittsburgh firm Three Rivers 3D.
The scanner uses two cameras to record a subject, measuring distance from the machine to create a precise map of the surface. The cameras can capture around 1 million points in only five seconds. We employed two machines, each with a different resolution and scale, the smallest measuring data points approximately every 80 microns (8,000 points per cm²).
The scanners feature a fully-automated turntable for 360-degree recording, but in many instances the artifact shape and size necessitated manual manipulation and recording. The more complex the surface topography of the artifact, the greater number of images required. In some cases, a single object dictated 30-40 individual scans.
After recording surface shape and texture, the individual scans were aligned and processed using the open-source software MeshLab. Although algorithms are available to automatically glue the scans together, frequently they blur and distort the data to an unacceptable degree, no longer preserving meticulous records of the original surface. Therefore, in most cases we manually glued the scans together in order to maintain the 3D model's value for scientific documentation and research purposes.
Manual alignment necessitates lengthy post-processing alignment time, and the files can be massive. We publish the 3D models in PLY format. Following the recommendation of Hörr and Brunnett (2013), our published models confine the number of vertices to around 4,000,000 or less.
Jackie Patterson employs the automated turntable to record a terracotta offering.
Jessica Marinaro aligns and processes an inscription using MeshLab.
Hörr, Christian, and Guido Brunnett. 2013. "Boon and Bane of High Resolutions in 3D Cultural Heritage Documentation." in H. Bock, W. Jäger, and M. Winckler (eds.), Scientific Computing and Cultural Heritage. Springer Science & Business Media, Berlin, 31-39.