It's getting hot in here...
As of today, the Zagora Infrared Photogrammetry Project is over. Despite the field season being over, I still have several lab notes I want to post on this forum. I always envisaged posting most of my lab notes after the season was complete and once we had results flying in.
The purpose of my Experiment.com fundraising was to purchase a FLIR VUE PRO R camera. This is an infrared camera, otherwise known as a Thermal Camera. In order to understand how this works for archaeology, let me copy a bit of an article of mine that will be published in the next month or two.
“Thermography, more commonly referred to as Thermal Imaging, is the practice of translating infrared radiation into a pictorial representation of heat (Meola & Carlomagno, 2003). Every object that is hotter than absolute zero emits infrared radiation, with the amount of radiation released increasing with temperature (Gaussorgues, 1994). A thermal camera can assess the amount of infrared radiation being emitted and translate that into a thermogram, more commonly known as a thermal image. The principle of archaeological thermography is that as the sun rises and sets during the course of the diurnal cycle, subsurface remains will absorb and emit infrared radiation, with the amount of infrared transference dependent on variables such as moisture, material and density (Perisset & Tabbagh, 1981; Haley et al., 2002; Casana et al., 2014). If the emitted thermal radiation differs from that of soil surrounding the feature, it may be detected by a thermal camera (Kvamme, 2008a).”
So, hopefully you get the idea of why this form of technology works for archaeology. In practice, even getting to the point of taking these photographs can long.
Firstly, we must set up ground control points. These were made out of cardboard and they are visible in both the regular photographic wavelength and also in infrared, as they will provide a marked difference when compared to the ground. We place between 3-12 of these around an area of interest, which would be anywhere from 80sqm to 8000sqm.
These were shot in using a GPS device using RTK services. Basically, this type of GPS is much more accurate than your average phone or car GPS. The device connects to satellites like a regular GPS, but then also communicates via a mobile phone to a fixed center. The location of this center is known to some amazing degree of accuracy. As we take points, the GPS asks the RTK Center for corrections in real time. This means we can get accuracy to around 1-3cm in three dimensions.
Once all of the GCP’s are shot in, we then fly the area with a DJI Phantom 4 Pro drone. This drone takes normal, everyday kind of photographs. We fly this in transects across the area, taking photographs every 5 seconds. These photographs are imported into a program called Agisoft Photoscan. This amazing piece of technology uses a special technique called 'Structure From Motion' to work out the 3 dimensional nature of the ground. It is tough to explain, so I might dedicate a post about it. But one of the outputs of this software is called an orthophotograph. Once again, without getting into too much detail, an orthophotograph is a photo that has all distortion removed, which makes it like a map of the area. Here is one of the south of the site.
Finally, we have to wait for it to get dark. Materials can reflect Infrared radiation. For example, a shiny surface will appear a lot colder than a rough surface, even if they are physically the same temperature, because the shiny surface will be reflecting the radiation from the sky, which is cold. So it’s best to shoot thermal imagery at night when that won’t be an issue.
We fly transects over the area we marked out, with the thermal camera shooting every 2-3 seconds. We then import them into a thermal imagery software and process them. We then stitch them all together into an orthophotograph as we did with the regular photographs. Here is a thermal orthophotograph of roughly the same area as the one above.
So, the long and short of it, we now have two big images. One from regular photography and one from thermal photography. We then compare the two on the computer. You are basically looking for things in the thermal image that doesn’t have a clear reason for existing based on the regular orthophotograph. Looking at the above image, you may make out a some lines of brighter(hotter) areas. What you are actually looking at is a large house or group of houses. If you join the large hot areas, you can see straight lines that make up the walls of this building. Sections of this area have been previously identified through Geophysical Survey, but our technique is a lot more detailed. For example, look at this small section of the building.
The two images you can see are the same. What you can hopefully see is that there is a roughly square shape(dark) with a brighter square around it. I have traced out the rough lines on the image on the right. What we have here is a room. The internal, dark lines is the inside area of the room and the white(hot) areas are the collapsed walls.
In the end, we have taken somewhere between 20-30 flights over areas and shot them with thermal. We have over 25, 000 photographs which we need to go through. It is a lot of work. But the results, which I hope you can see above, are amazing. All thanks to you and the generosity of all the people who donated to this project.
1 comment