Digital Elevation Model - lower areas are blue

How big is that hole? Drones investigate..

Peter Rauxloh
07.12.2016

Our Director of Technological Solutions, Dr Peter Rauxloh, explains the importance of using drones to accurately measure and monitor in archaeology.

Archaeologists are of course very interested in digging holes, and carefully recording the size of those holes and the features they discover within them.  Keeping an eye on how much excavation has occurred, is as important for us as it is for the people we work for, since it provides information on the rate of work,  how much has been done,  how much is left and how much post-excavation work is likely.  An archaeological project, like any business, has to balance it’s resources,  so the project may be properly completed to the satisfaction of all.

The question of just how much excavation has been carried out, can be answered in a number of ways:

  • A site will have a foot print and the rough dimensions of that foot print will provide us with a an area.
  • If we then multiply the area by the average depth, we obtain a volume – primary school mathematics is great!

However,  things get more complicated when the shape of the area is more complex, and the bottom of the hole is anything but flat.  For example, if we wish to determine the volume of material that has been removed by erosion  - perhaps  a coastal  site  that is nibbled by the sea – we really need to be more precise.

In these situations we need to consider two surfaces; a before surface and an after surface.  If you imagine a box with a lid, the before surface is the top of the lid, the after surface is the bottom of the box,  after the lid has been removed.  If that lid was covered with a grid of squares, you could ask of each square,  how far is it  above the bottom of the box? By adding up the answers from every square and multiplying by the size of a square, we get a volume.  This simple principle also works where the sides of our box are not straight,  and neither the lid or bottom of the box, is flat. The challenge is to accurately create these two surfaces.

A recent project poised just such a challenge and we deployed drone MOLA-SUA-2 to investigate.

On the remarkable low-lying landscape of Orford Ness in Suffolk, now under the careful stewardship of  the National Trust, work was underway to repair bank defences damaged after a storm surge, by removing material from dry land inside the banks and using it to build a new bank.  Orford Ness is a shingle spit which joins the mainland at Aldeburgh and was formed by the action of the sea, known as long-shore drift.  It runs parallel with the coast,  separated from it by  the river Alde/Ore,  down to Orford and finally North Weir point.  The site  is an internationally important one  for nature conservation,  and has a fascinating history of  occupation, most vividly,  it’s role as a military testing site for early experiments in fighting aircraft, radar and atomic weapon research. 

The subjects of interest were two large but shallow excavations roughly the area of a football pitch. Our work followed the normal process of laying out and carefully measuring the position of some 9 ground targets, and then flying the aircraft over them in a tight pattern like a tractor ploughing a field. This process is known as aerial survey. Indeed, since we were keen to accurately capture the height of the surface (rather than just it’s area) we flew the aircraft across the field too, (i.e. like ploughing up and down the field and then  from left to right – not advisable on actual fields in actual tractors!). 

The 317 images captured by the aircraft,  in combination with the ground target’s coordinates, were processed to produce a single accurately scaled and positioned 2D image of the site.  This orthmosaic, as it is known, is a single image made up from those 317 images such that one is looking directly down over  the entire area -  a view that any individual image cannot provide.  As well as this orthomosaic , we produced a second plan of the area, in which each single cell or square,  indicates a height.  This Digital Elevation Model (or DEM)  is our after surface.   Before the excavation commenced,  a traditional survey had been carried out over the area, which  provided   some 110 3D points from which we were able to produce our before surface.

Having carefully traced round the edge of the excavation, so we may consider just this part of the two surfaces, we then subtracted the after surface from the before surface – just as in the example of the box with its lid.  

In this way we were able to make a very accurate calculation of the material removed,  to help the National Trust with the monitoring and management of this important site.  The use of the aircraft meant that we captured the data accurately and rapidly,  the flight taking less than 15 minutes. The comparison between the traditional survey that took place before the excavation and the flight is perhaps the most important aspect.  The former provided a 3D point every 15m or so,  on which we based our surface,  and probably took an hour to complete.  The latter was completed in a fraction of the time allowing a number of other jobs to be undertaken, but most importantly,  it provided 73 3D points per square meter; a density of coverage more that two orders of magnitude greater.

 

  • From the experts
  • Technology
  • Science
  • Research
  • Excavation

Related blogs