My name is Piotr Leżański, I’m an archaeologist from Gdańsk. Although my Sketchfab account doesn’t reflect it (yet), my main area of interest is spatial analysis of lithic artefact distribution on Mesolithic sites, focused on both: reconstruction of hunters-gatherers’ camps internal organization and identification of post-depositional processes and how they affect relics unearthed during excavation. Beside this, for almost 12 years now, I have been working for various institutions, museums and private companies in preventive archaeology on prehistoric and medieval sites endangered by planned and ongoing construction. One of the common things for both lines of my work is attempting to make use of the progressing informatization of archaeology. Needless to say, photogrammetry and 3D scanning are very important elements of this process.
I don’t remember when exactly I first came across photogrammetry, but I remember being impressed by the results of using this method to scan lithic artefacts presented in 2016 paper “A Simple Photogrammetry Rig for the Reliable Creation of 3D Artifact Models in the Field: Lithic Examples from the Early Upper Paleolithic Sequence of Les Cottés (France)” by Samantha Thi Porter, Morgan Roussel and Marie Soressi. At the time I was more focused on analysis of dense flint scatters, so I treated it as something interesting and worth knowing, but not very useful for what I was working on. Later, however, getting familiar with photogrammetry allowed me to use it, to a greater or lesser extent, during rescue excavations, as it can provide accurate field documentation in a short time—what in most cases is very important due to tight deadlines. After some time I’ve returned to the subject of using close range photogrammetry for creating 3D models of artefacts and it has become part of my workflow during post-excavation work.
Currently, when some material is handed to me for analysis and scientific description, I usually pick some more interesting pieces and capture them in 3D. Besides the analytical value that this step can provide, one of the reasons to do this is the fact that the vast majority of these materials will never be exhibited in any museum.
Fortunately, the companies that I work for are usually enthusiastic about showcasing models of artefacts on platforms like Sketchfab. Thanks to that they can be used to promote cultural heritage or as a reference for other scientists. There’s also great educational potential—obviously 3D models will never replace the ability to touch and to see an artefact in person, but, to an inexperienced archaeology student, it can provide better perception then drawing or photography. Finally, I feel obligated to create the best possible documentation, with the means available to me, for any future research.
When capturing single artefacts, workflow doesn’t differ from what is presented in many other places in the internet or literature. I use a uniform background, turntable, and DSLR camera on a tripod. A single LED panel with neutral white colour, mounted on top of the camera, is usually enough for lighting. The photographed item is placed on a sheet of graph paper. Although it’s usually not a problem, sometimes I put some additional marks on this sheet to break up the repeating pattern. This piece of paper has two functions: to help with photos aligning and to serve as a scale. The strategy for taking photos depends on the object’s shape. For objects with enough volume (like fully preserved vessels or lithic cores), photos are taken from 3 heights, then the object is rotated by 180 degrees and whole process is repeated. This approach doesn’t work very well for small and thin objects like lithic blades and potsherds. One thing is that it produces more photos than needed, and another thing is that, after the graph paper and any other objects used for immobilizing the item are masked, there’s usually a problem with aligning front and back, due to insufficient overlap on photos taken from the sides. In this case, I make 4 sets of photos—the object is rotated by 90 degrees and photos are taken from 1 or 2 heights. If it’s still not enough, in the photogrammetry software I leave one set of photos unmasked and the graph paper is used to align the front and back of the item, although I try to avoid this approach, since there is more cleaning up required later.
For processing photos, I use Agisoft Metashape standard. Two aligned models are created: the final version without any helpers and the model made from a single unmasked set of photos. The latter is used to set the proper scale in Meshlab. Finally, Blender is used to render images that can be added to the documentation and used in a publication.
In the case of models that are later uploaded to Sketchfab or shared via other media, I also take some time to set PBR materials. Besides the aesthetic aspect, the aim of this step is to at least approximately imitate the surface of the item, as it is one of the attributes included in technological analysis of prehistoric pottery. For objects with a rather uniform surface, it’s usually enough to use specular and roughness sliders in the Sketchfab viewer, but if there is more variability I make use of the fact that different roughness is, in most cases, correlated with differences in colour. In Substance Designer I use base colour texture as input, ‘color to mask’ nodes to set roughness value for separate areas and then blend them together to get the final texture.
The vast majority of artefacts gathered during excavations are strongly fragmented and, for that reason, archaeologists use various techniques of reconstruction to provide better perception—from illustrating estimated shape and size on schematic drawings, up to using clay, gypsum, or similar material to make a physical reconstruction of the artefact. Bringing an artefact into virtual space provides more options on this subject—creating digital reconstruction is one of them.
Once again, workflow strongly depends on size and shape of the original piece. Sometimes it’s necessary to go over literature to find images of similar items and use them as a reference for creating models almost from scratch, but sometimes there’s also enough of the original artifact’s geometry preserved to be able to use only it (especially when it comes to the pottery). In this case, after finding the center of radius, the scanned model can be duplicated and rotated several times to create a base mesh used for retopology. After the reconstructed model is finished and UV unwrapped, Substance Designer is used to bake colour texture, and normal map from the scanned model; later the clone brush in Substance Painter is used to paint both of them over the rest of the model.
After seeing some examples of successful implementation of photogrammetry to read poorly visible stone engravings, I’ve started to explore the possibilities of using similar methods to analyse surfaces of scanned potsherds with a goal to identify ornaments on strongly weathered and damaged pieces. Currently I’m at the early and very experimental stage, so I don’t have any clearly established workflow that can be described here, but after my first more or less successful attempts, I’m able to say that relief extraction looks like a very practical application of photogrammetry in archaeology.
Favourite 3D scan
Sketchfab gives an opportunity to see various interesting objects from all over the world, from impressive constructions (for example these presented by the Zamani Project), to interesting artefacts from distant (to me) places (for example). There are also scans of artefacts from important sites, like the paleolithic site in Kostenki, that I know from the literature. But if I have to pick one favourite, I would go with the Jericho Skull, because:
- The artefact itself is very interesting. It comes from an early stage of neolithization—a process that began our current way of life (by abandoning hunter-gatherer economy and nomadic lifestyle in favor of farming and sedentary lifestyle), yet it illustrates a big cultural difference between the past and today;
- It’s a high quality model presented in a very attractive and informative way.