Three-dimensional model of a Neolithic juvenile burial. (Click to view in 3D)

This week I want to write about changes in the way we’re recording burials at Çatalhöyük. Previously, once a burial was fully excavated and cleaned up for a photo we would draw the skeleton and grave cut in plan view at a 1:10 scale (1:5 for infants). This is standard archaeological practice on most excavations; the plans are typically used in conjunction with photographs in order to provide a visualization of the orientation of the body within the grave and any grave goods that may be associated with the burial. A problem with planning burials by hand, however, is that the accuracy of the scale drawing depends on the skill level of the archaeologist.

While some drawings are technically – as well as artistically – rendered and highly accurate in terms of scale, others are little more than stick figures that bear little relation to the actual burial besides the orientation of the arms and legs. Another problem is the amount of time it takes to produce the drawing. Setting up a planning grid and preparing the drawing board takes time, while actually drawing the skeleton itself can last several hours.The longer a skeleton is exposed to the elements the sooner it will begin to deteriorate. As such, reducing the amount of time a skeleton is exposed to the environment will contribute to the long term preservation of the bone. One way of doing this is to photograph the skeleton in plan view and trace it digitally using AutoCAD or ArcGIS. In this way the skeleton can be lifted and taken to the lab for processing as soon as it’s been photographed and the archaeologist can produce the digitized plan at a later date. This year we’ve finally taken the plunge and moved towards a fully digital workflow for producing scaled plans of the burials. We’ve also begun experimenting with three-dimensional modelling of burials in order to help us work out the often complicated stratigraphic sequence of interments under the Neolithic house floors.

Unlike 3D laser scanning, which requires very expensive, specialized equipment to produce high quality models, the methods we’re employing require no more than a point and shoot camera and two pieces of software: PhotoScan and MeshLab. Nicolò Dell’Unto, from Lund University in Sweden, has been testing this system at Çatalhöyük since 2011 and this year he’s taught me to use these methods for recording burials. The procedure requires taking a series of photographs of the burial from as many angles as possible. Depending on the nature of the burial, this can be anywhere from fifteen to fifty photographs.

3D model of a burial showing the number of photos used and the angle from which each was taken

The three-dimensional geometry of the burial is generated from the photographs to form a polygon mesh which is then overlain with a photo-realistic texture produced from the images themselves. These models can then be scaled and georeferenced for placement in Geographic Information Systems (GIS). Determining the order of multiple interments under house floors can be extremely difficult, but with this technology we can virtually reconstruct the sequence of burials by stacking together a series of georeferenced 3D models produced during excavations. Using this method we were able to make sense of a very complicated set of burial sequences in the North Area of the site and also to identify a possible skull retrieval pit.

Stacked 3D models in MeshLab

Finally, the 3D models can also be used to produce orthogonal, plan view photos of burials which can then be digitally traced and imported into the project GIS. Unlike a single photo taken from above the burial – usually with a wide angle lens – these orthoplan photos are less susceptible to planar and lens distortions and thus easier to rectify and georeference. The only drawback at the moment is the computer processing power required to generate and stack the 3D models. The project will require several dedicated, high-powered desktop computers to carry out the work.

Orthogonal plan view of a Roman double burial generated from a three-dimensional model.

UPDATE: My colleagues and I presented a poster (below) based on this work at the 2013 Society for American Archaeology (SAA) conference in Hawaii, as well as the 2013 American Association of Physical Anthropologists (AAPA) conference in Knoxville, TN. Click on the image below to view the full-size poster.