SUAVe: Redefining Archaeology With Aerial AI
There are limits to how fast you can complete a survey of an archaeological site or dig Due to the fragile nature of long-buried ruins and remains, a survey by hand can take years two to three years on average according to archaeology expert Steven Wernke who works at the University of Vanderbilt as a dig archaeologist.
Even AR scanning tools can only speed things up so much,. By the very nature of a dig, it is not flat terrain, and any little mound of dirt or shard of pottery is going act as a barrier to the scanning beams of a SLAM system, and build up an incomplete picture. The external-scanner is going to have to be moved many, many times to build up a complete picture something that is barely faster than doing it by hand.
However, since the terrain itself is a problem, perhaps the solution is to get away from the terrain entirely? An aerial scanner perhaps? That's certainly the intent behind the SUAVe system. SUAVe stands for Semi-autonomous Unmanned Aerial Vehicle, and is a new hybrid UAV jointly developed by Wernke, and another expert at the university; engineering professor Julie Adams.
The SUAVe (pronounced SWAH-vey) system should transform how we map large sites that take several seasons to document using traditional methods. It will provide much higher resolution imagery than even the best satellite imagery, and it will produce a detailed three-dimensional model, Wernke said.
Because the aircraft takes pictures continually as it flies, the problem of occluding terrain or overhangs is dealt with. Whilst a particular feature may hide other features behind it when seen from a particular angle or elevation, it won't hide those features from all angles and elevations. So, as the plane flies overhead, taking pictures on pass after pass, angle after angle, it collects all the data a researcher could ask for, in order to reconstruct the scene in 3D.
The UAV doesn't handle this processing itself. Instead it dumps all the data it collects into a ground-module which performs all the processing necessary to stitch the photos together and create an interactive 3D landscape from the data. This deals with two important points. Firstly, it keeps the computing equipment on the plane to an absolute minimum, alleviating weight concerns. Secondly, since the ground station doesn't even have to be at the dig site, it means that the archaeologists aren't limited to bringing a suitably trained technician along with them. Anyone can launch the plane and direct it as to what area to fly over and photograph, or when to come in for an early landing (and where). This makes it much more accessible for research projects of any size and scope. It also opens up the possibility of using the plane in truly remote areas with little or no support.
You will unpack it, specify the area that you need it to cover and then launch it, Wernke said. When it completes capturing the images, it lands and the images are downloaded, matched into a large mosaic, and transformed into a map.
The algorithms developed for the project allow the SUAVe system to specify the flight pattern to compensate for factors such as the wind speed, the angle of the sun and photographic details like image overlap and image resolution, Adams stated.
The first tests of the SUAVe system are already under way. Running from mid July 2012 till mid August 2012, they are being conducted at a dig site at the abandoned colonial era town of Mawchu Llacta in Peru. This will be used to iron out any issues with the aircraft in actual dig conditions over an extended period of time. The team plan to return next year to try again, if anything does require tweaking.
Archaeology is a spatial discipline, Wernke said. We depend on accurate documentation of not just what artefacts were used in a given time period, but how they were used in their cultural context. In this sense, SUAVe can provide a fundamental tool set of wide significance in archaeological research.
Wernke hopes that the new technology will allow many archaeological sites to be catalogued very quickly, since many are being wiped away by development and time.
The SUAVe system should be a way to create a digital archival registry of archaeological sites before its too late, he said. It will likely create the far more positive problem of having so much data that it will take some time go through it all properly.