The Discovery Programme has had a long standing research interest at the Hill of Tara, Co Meath, one of Ireland's best known archaeological sites, undertakin g extensive topographic and geophysical survey on the ground since 1992. The terrain model generated from total station survey in the mid 1990's was ground breaking at the time, a significant factor in furthering the understanding of the archaeological remains. Our desire to increase both the extent and resolution of the terrain model was realised when the Heritage Council provided financial support to enable a high resolution LiDAR to be commissioned.
The objective was to generate both a Digital Surface Model (DSM) – the landscape including trees, hedges, buildings etc (known as the first return) and a Digital Terrain Model (DTM) – the landscape with all upstanding features filtered out (known as the last return or bare earth). Both models would have a ground resolution appropriate to enable even the most subtle elements of the micro-topography to be seen, but this would not be restricted to simply the known monuments. Our intention was to cover the entire Tara landscape at this resolution, potentially revealing new discoveries, enhancing the understanding of the wider area and providing a definitive topographic base map, a GIS resource, to which further scientific data such as geophysical survey could be added.
Unlike most of the survey projects under taken by the Discovery Programme the primary data was captured by a third party, we do not possess our own airborne LiDAR system. Instead we went to a LiDAR data provider (Fugro BKS) and designed a survey specification which would supply us with two sets of x,y,z data from which our DSM and DTM could be generated. A target resolution of 60 pts per metre2 was set – a notional ground spacing of 12.5cm.
Instruments and software
The LiDAR system used was FLI-MAP 400, mounted beneath the fuselage of a helicopter. The sensor system consists of three 150kHz LiDAR sensors, one 15o forward, one nadir and one 15o aft, two RTK GPS receiver, an inertial navigation system (INS), an 11 megapixel digital imaging sensor and a digital video feed.
Substantial data processing using FLIP7 software was completed by Fugro-BKS before the ascii data sets were supplied to the Discovery Programme. These included: transforming the data to WGS84 & Irish Grid coordinates; production of the tiled ascii DSM data set; removal of vegetation, buildings and above surface features using a combination of intensity and video inspection; and finally production of the tiled ascii DTM dataset.
Both DSM and DTM ascii datasets had to be tiled as the total point number in each was unmanageable - 150 million pts. Surface models were created using ESRI ArcGIS 9.2 utilising the 3D Analyst and Spatial Analyst extensions. Firstly TIN models were created from each tile ASCII xyz data set. These were subsequently rasterised to improve display performance and merged into a single seamless DTM and DSM. Hillshade models of the DSM and DTM surfaces were then generated to give the spectacularly detailed GIS data layers which are the basis for advanced archaeological research on the site.
To extend the DTM for the Hill of Tara required a different approach than ground survey. Even with access to improved technologies such as robotic total stations or RTK GPS it would have been prohibitively expensive to continue modelling by ground survey. We had some experience generating DEM's through digital photogrammetric processing but this had made us aware of some of the limitations of the technique – particularly the fact that vegetation such as trees and hedgerows are included in the model, and the difficulty generating DEMs at the high resolution required. Our experience with fixed wing LiDAR – with resolution between 0.5 – 1m – had generated spectacular models, identifying archaeological elements within the wider landscape, but failed to give the micro-topographic detail we sought for the Hill of Tara. For these reasons FLI-MAP LiDAR appeared to offer the ideal solution.
The results appear to vindicate our decision to use this technique. As we haven't applied fixed wing LiDAR to this site we have no direct comparison but by comparing models from a neighbouring site, Bru na Boinne World Heritage Site, we believe the case for FLI-MAP is compelling, see illustration.
The major problems in the initial processing of the data were due to the volume of data being handled. To process the data in GIS and generate our DEM and DTM required the data to be tiled, but our output deliverables had to be seamless complete elevation and hillshade models. This issue was resolved by creating tiles a pixel beyond the original boundary and then averaging the overlap values when merging the tiles into the final model.
The hillshade models were initially generated using the default ArcGIS Hillshade function angles (Azimuth 315, Altitude 45) but early examination showed the limitation in simply using one illumination angle. Features could be hidden or visually suppressed due to their aspect rather than size. This was resolved by generating hillshade models based upon multiple light sources correlated to the frequency of relief features.
The primary deliverables were the GIS products; the DTM, DSM, and associated hillshade models. These are available in the GRID format – georeferenced image files with a 0.1m cell size. Also supplied are the 100m x 100m orthoimage tiles in ECW format and the AVI files from the forward and nadir video feeds.
The GIS products (and the GIS compatible orthoimages) have been extremely well received by the archaeologists involved in research at the Hill of Tara. With the data in such commonly used formats they have been easy to distribute to colleagues and other researchers and have been used to reveal new discoveries and enhanced interpretation of existing sites, in particular when interrogated in conjunction with geophysical survey data. From a project planning perspective they have proved a valuable resource in determining areas suitable for geophysical prospection.
This approach to landscape mapping has become much sought after by archaeological and heritage agencies in Ireland with a number of projects following this level of specification. Cost is the major determining factor but the value in terms of research and heritage management has been recognised. From our perspective our only reservation would be the limited area we were able to cover by this technique given the funding available. It again creates an artificial boundary around the Hill of Tara archaeological area at a time when research is beginning to see the hill itself as part of a much wider landscape.
Anthony Corns, Robert Shaw, High resolution 3-dimensional documentation of archaeological monuments & landscapes using airborne LiDAR, Journal of Cultural Heritage, Volume 10, Supplement 1, ICT and Remote sensing for Cultural Resource Management and Documentation, December 2009, Pages e72-e77, ISSN 1296-2074, DOI: 10.1016/j.culher.2009.09.003
Corns A., Fenwick J., and Shaw, R.,2008. 'More than meets the eye', Archaeology Ireland, Vol. 22, No. 3, Issue No. 85, pp 34-8