The Potential of Airborne Lidar for Detection of Archaeological Features under Woodland Canopies
Devereux, B. J., Amable, G. S., Crow, P., Cliff, A. D., Antiquity
For the last half century vertical and oblique aerial photography has been the mainstay of airborne archaeological reconnaissance. But, despite its immense success it does have certain limitations. In particular, lighting conditions and sun elevation must be at or near optimum levels to detect subtle crop marks. Moreover, where archaeological features are hidden beneath woodland canopies they cannot be seen. This paper highlights the potential of high resolution, digital elevation models derived from airborne laser scanning (lidar) to overcome these problems. It demonstrates that when lidar is used in conjunction with hill-shading techniques that enable variation in the direction and elevation of the illumination source, more effective detection of features is possible. Furthermore, it shows that so called vegetation removal algorithms have the capacity to reveal features hidden from aerial photography by woodland canopies in considerable detail. The prehistoric hillfort at Welshbury Hill, Gloucestershire, is used as an example. It is concluded that airborne lidar has enormous potential to extend archaeological reconnaissance into new and exciting areas.
Limitations of conventional aerial photography
The detection and mapping of archaeological sites using aerial photography is one of the oldest and most successful branches of remote sensing. Photographs of monuments such as Stonehenge in the United Kingdom date back as far as 1906 and were collected with hand held cameras from army war balloons (see Wilson 2000). With the deployment of vertical cameras in fixed wing aircraft for military reconnaissance in the 1920s it soon became evident that aerial photography had vast potential for archaeological applications. This potential was realised by O.G.S. Crawford, the first Archaeological Officer of the Ordnance Survey, who demonstrated the effectiveness of the technique for mapping earthworks in Wessex (Crawford & Keiller 1928). The rapid growth of aerial survey information and the exciting consequences for archaeological research prompted Crawford to establish the journal Antiquity in 1927 (Wilson 2000).
The particular feature of aerial photographs which makes them so valuable for aerial reconnaissance is their ability to reveal buried features via crop marks at a scale relevant to the historic landscape. Coupled with its capacity for photographing large areas at a very high spatial resolution it is not surprising that aerial photography has become a central tool for archaeological mapping. Amongst others, the University of Cambridge has conducted ongoing aerial survey programmes since the 1950s aimed at detecting and mapping archaeological sites. The photography resulting from these activities has been central to studies in landscape archaeology (St Joseph 1966). It has resulted in major collections of photography housed within the University of Cambridge (www.uflm.cam.ac.uk) and within government agencies charged with heritage management. The English Heritage National Mapping Programme is heavily reliant on aerial photography from such resources (Bewley 2003).
The advent of colour photography and photogrammetric techniques has been important in strengthening the capacity of aerial photograph interpretation for revealing archaeological information (Mead 1983, Burnside 1985). Ultimately, however, success depends on the photography containing crop marks which correspond with buried features. Crop marks occur for two main reasons. Firstly, buried features such as ditches cause very slight hollows to occur and equally slight mounds over ramparts, walls and buildings. If photographs are collected when the sun is low in the sky these shallow earthworks cast shadows within the crop which can be detected by the photo-interpreter. Secondly, deeper soils over in-filled ditches and shallower soils over solid features result in variability of moisture regimes and availability. …