Wednesday, December 11, 2013

Call for Participants and 3D computer modelling of Oven Site

Smiths Island seems a long way off and a long time ago, considering the view of snow-covered trees currently out my window, but already I am putting together next season's excavation plan and finishing the 2013 season site report.  This year, the University of Rochester will be opening up participation in the field school internationally. Participants can earn six college credits (undergraduate) in Bermuda actually for less than the price of taking the same course in Rochester, NY, and this includes food and lodging for five weeks in Bermuda, under my expert tutelage ;-)  Graduate credit can also be arranged.

Those interested can download an application at my my department homepage.

The deadline for applications is March 1, but I will be filling spots on a rolling basis before then, so you are encouraged to apply early. Please contact me (michael.jarvis@rochester.edu) for more information or answers to specific questions.



I don't know if any of you wonder what goes on in the ten months between the frantic flurry of postings on this blog, but rest assured I'm not sitting on a sofa eating bon-bons. Lots of teaching and working with students on analyzing the Smiths Island sites accounts for a lot of the off season - in general terms, for every hour spent in the field, archaeologists devote four to artifact processing, analysis, record digitization, GIS input, writing reports, etc. In my case I cannot work with the site assemblages off season, since by law all artifacts must remain in Bermuda and I'm in (now snowy) Western New York, but recent advances is digital technology and my university's acquisition of a GIS specialist and ArcGIS license has facilitated highly detailed virtual reconstructions of our sites and the linking of a wide range of historical records spatially to Smiths Island. Anima and Leigh have also both been quite busy as well in the months since the dig making this happen!

Last week the University of Rochester hosted a symposium, 3D Digital Archaeology: Reconstruction, Analysis and Conservation of Cultural Heritage, which brought together scholars working on the Roman World and pre-Columbian Peru in support of U of R's Archaeology, Technology, and Historic Structures program (Thanks, Renato and Elizabeth!!!). It was an incredibly useful gathering for learning about cutting edge technologies, seeing them modeled, and thinking about the new insights and modes of analysis they generate. I was especially interested in several software programs/services (PhotoScan, Pix4UAV, ReCap) that process photographs into manipulable 3D models and the prospect of using UAVs (unmanned aerial vehicles, or Drones) with HD cameras to capture whole landscapes in 3D, exportable into GIS.

After a few days' cogitations, I wondered if I had inadvertently captured enough photos of Bermudian sites and the Smiths island sites during the 2012 and 2013 seasons for one of the programs (Autodesk's ReCap, in this case) to process. My first attempt was to upload several dozen reference pictures I had made of the early 18th c. ruin called Longford in St. George's off Duke of Kent Street. After a few missteps and a couple hours' wait, an impressive model arrived in my inbox:
This picture does not do justice to the awesomeness of this model, which can be orbited, rotated, and pivoted and has a CAD web of about a million datapoints underlying the draping of the stitched together photographs. Not bad, considering I did not shoot any of the photos with the proper method of extensive overlap, multiple heights and angles, and including common reference points for ease of modelling.

This success led me to invest many more hours selecting and uploading nearly 200 Oven Site photos. Again, having not taken these photos with 3D purposes in mind and with insufficient overlap, I didn't expect much. I spent four hours last night uploading them into ReCap and then went to bed, letting Autodesk's computers work through the night. 

The model that greeted me this morning went beyond my wildest expectations - and these images are only coarse graphic projections: the program also generates sub-centimeter versions downloadable into AutoCAD 3D and other platforms.
A photomosaic of Oven Site from the top, showing aggregate 2010, 2012, and 2013 excavations (facing North)
And now orbited and tilted to show northern sidewalls
Further rotation, facing east: photographs of the uncut bedrock surface at the far eastern edge were insufficiently cross-stitched to be displayed.
The site rotated to give a view looking west. The gray and white areas in the foreground lack sufficient overlapping coverage for modelling.
The program also can reveal the CAD datapoints underlying the visualization thus (same view as above):
And also display the camera positions and planes of the individual photographs amalgamated into the model (the gray squares):

The conference, in short, was transformative for me in unlocking a lot of the potential of data I'd already collected. Now, in my cold Rochester office, I can re-explore the Oven Site, take measurements, project proportions, and revisit interpretations and assumptions made during the field season. More profoundly, the model is really 4D, in that it collapses time: the integrated photos were taken in 2010, 2012 and 2013, effectively creating a visualization that has never existed but uses details and data collected diachronously.

I can't wait to return to Bermuda for the 2014 season with these tech tools in mind so I can capture the proper array of overlapping multi-perspective photos to create more complete and fine-tuned 3D arrays, both on Smiths Island and in the Town of St. George's - after all, standing buildings work as well, if not better, than ruins and archaeology sites.

Anyone want to buy me a UAV and a GoPro Black for Christmas?

1 comment:

Virginia Bernhard said...

Bravo, Jarvis!

Such good work!

Keep it up!