The increasing sophistication of digital survey equipment over the last 20 years has enabled us to apply new approaches to data collection that both enhance our data and its potential for increasingly sophisticated spatial analysis. The Total Station (TST), based on solid 19th-century theodolite technology, is now a standard piece of equipment on excavations, providing the basis for the capture of high-precision digital survey data. The evolution of these devices from theodolites with a top-mounted Electronic Distance Meter (EDM) and separate data logger to the all-in-one modern TST, which can be operated in reflectorless and robotic modes, have revolutionised the ease with which surveys can be undertaken.
During the years since 1984 when the first EDM was used in Heslerton, more than a million 3-D points have been collected, the majority providing 3-D locations for individual finds, mostly animal bone (see above). The LRC use of the TST has a bearing on all aspects of the record. It is, for instance, standard practice in most excavations to lay out a physical grid with pegs at 5m or 10m intervals. In excavations where individual trenches may cover over a hectare and cover areas of bedrock where installing the grid pegs accurately is difficult, the process is very time-consuming. An alternative approach is to establish a number of base stations around the excavation, located on the Ordnance Survey national grid, and create grid-points using the TST as and when required. Planning sheets need not be aligned on the site grid but can be located and orientated for the convenience of drawing the plan and need only incorporate, at a minimum, two surveyed points to ensure that they can be accurately digitised in the correct location; this approach is particularly beneficial when recording widely distributed and isolated features.
Perhaps, since the process generates a hiatus in the excavation and it can take a considerable time, the preparation of field drawings is perceived, by some, as an aspect of the recording process that 'needs' a digital solution that will speed up the process. The use of TSTs to generate the primary site plans is a subject that initiated considerable discussion during the DigIT excavation. Digitised, as opposed to scanned, excavation drawings are the primary visual component used in GIS-based excavation data management and analysis. Being vector based, they can be viewed at any scale and, since they are linked to the excavation database, can be used to visually display the results of any SQL queries applied to the database or to reflect the sequence documented in the stratigraphic matrix. They can also be used to generate publication drawings at any scale without the need for any redrawing, as would have been the convention in the past.
If excavation plans were merely pictures, then we could simply record them through rectified photographs, thus recording a great deal of visual detail in a matter of moments. However, excavation plans are far more than pictures and part of their importance derives from the process of creating them. Archaeological features (Contexts) are complex three-dimensional surfaces, whether they be excavated pits, walls or layers of soil. The majority are either poorly defined or defined by the trowel of the excavator, and the surface may include cultural debris and materials imported to the site, such as building materials. In almost all cases there is a considerable degree of interpretation in their definition; most features 'belong' with others such as the single post-hole in a wall line.
In addition to the complex 3-D surface that we try to record through a line drawing, using conventions such as hachures, aspects such as the colour, and particularly texture, of the soil are also important. Hachures are designed to indicate slope shape on two-dimensional drawings, and do not easily convert to digital format, and in the case of excavations undertaken in Heslerton have been avoided in favour of contours drawn at 5cm intervals, which are digitised in 3-D.
The need for field drawings, both plans and sections, is defined on the basis of two criteria, the need to fulfil the fundamental requirements of any excavation – preservation by record and the articulation of interpretation. Since the process of excavation is, for the most part, entirely destructive, the physical resource being examined must be carefully recorded: the detailed recording cannot be overlooked in favour of a quick but approximate solution. Post-holes, for instance, could be simply 'planned' by recording a centre point, a diameter and a depth – all very well if one only wants a dot distribution map, but not much use if one wants to understand the structure they once supported. An even less useful approach to the post-hole recording exercise, is to record three or four points around the perimeter and use a smoothed curve to define the planform of the post-hole. This approach will produce a pretty plan and give the impression that the shape has been accurately recorded when it has not. The process of drawing excavation plans and sections takes time, not simply on account of the time taken to physically set up a planning frame and draw the object of investigation, but because it involves intensive study of the area to be drawn. This can expose new information not necessarily obvious during the physical excavation process.
During the DigIT excavations we were fortunate to have the assistance of Tom Cromwell, an experienced archaeological surveyor employed by English Heritage, who brought with him the most up-to-date robotic and reflectorless TST linked to a pen-based computer running mapping software. The singular advantage of using the TST to plot context outlines is the fact that the results are truly three-dimensional. The limitation of this technology is the time taken to record sufficient points to accurately describe the shape of the context being recorded. For overall low-resolution pre-excavation plans the TST provides the perfect solution as detail, at this stage, is less important. However, for documenting complex surfaces it is impossible to match the drawn record within the same timeframe, even allowing for the additional time required to trace the drawn results using a digitising tablet.
The difficulty of gathering high-resolution plans with the TST is not simply constrained by the time required to log each 3-D point but also by the difficulty of maintaining the eye-ground-plan contact required to follow a subtle edge. Used in reflectorless mode the field of view of the TST is too limited, and used robotically with a reflector staff, the need to keep the staff vertical and the point on what may be a very soft surface completely interrupts the flow of the drawing process. In the case of a manually drawn and digitised plan the individual points that describe an edge are spaced at sub-centimetre intervals, something not realistically possible with the TST. Even with a robotic TST recording automatically, drawing speeds at high precision are slow, and any errors in following an edge are very time-consuming to rectify; moreover the whole eye-ground-plan contact is interrupted and attention is drawn away from seeing and understanding the archaeology in favour of the technology.
It is likely that we will soon see intermediate solutions that combine the use of a TST or GPS with Tablet PCs, which could potentially produce accurate drawings at sufficient resolution. There are a number of technical and physical problems to overcome when using Tablet PCs; the relative fragility of the hardware, the slippery drawing surface, the limited functionality of the stylus and difficulties arising from the low screen resolution. Tablet PCs have high-resolution digitising hardware within them but harnessing this in an intuitive fashion so that millimetre precise drawings can be created on screen may require people to learn to draw in a way that is quite unlike the process on paper, where the combination of pencil angle, variable pressure and effective resolution reflect a sensitivity that is difficult to match on a hard and pixel-bound computer display.
The role of the TST for digital site planning must be determined on fit-for-purpose principles. While it is difficult and time-consuming to generate sufficiently detailed excavation plans using the TST, newer technology in the form of 3-D High-Definition Scanners (HDS) may provide an alternate approach. HDS instruments can accurately record 3-D surfaces at c. 2mm precision. High-precision scans do take some time to complete if large areas are to be covered, but with the newer scanners from Leica and other manufacturers now including digital cameras, the technology can generate what can realistically be described as high-resolution, accurate, three-dimensional photographs. These instruments are currently expensive, at about £80,000 with software and necessary computer hardware, but, within a few years the costs will radically reduce and second-hand models will become affordable for larger archaeological operations. Even HDS instruments will not fully solve the recording problem, as there will still be a need to identify the limits of the contexts being recorded and a comprehensive on-site experiment is needed to determine the ease with which this interpretative layer can be added to the model.
© Internet Archaeology/Author(s)
URL: http://intarch.ac.uk/journal/issue27/2/6.4.html
Last updated: Wed Nov 11 2009