Digital terrain modelling for archaeological interpretation within forested areas using full-waveform laserscanning

Abstract

The identification of sites within forested areas is one of the remaining unresolved issues for archaeological prospection. Airborne laser scanning can be a solution to this problem: due to the capability of penetrating forest to a certain degree (depending on the vegetation density) the determination of the terrain surface is even possible in wooded areas. To be able to identify archaeological structures, archaeologists have to interpret the resulting topographical data of a filtered ALS scan. This does not pose major problems with large structures. Smaller features, however, are much more difficult to identify, because their appearance in an ALS point cloud is very similar to natural and recent features, as for example dense brushwood, or piles of twigs or wood. Therefore, to eliminate potential sources of error, a high quality separation of terrain and off-terrain points is essential for archaeological interpretation while maintaining a high point density of the ALS data. Using conventional ALS systems, the possibilities to classify terrain and off-terrain points are limited and the results - especially in forested areas with dense understorey - are far from ideal for archaeological purposes. This paper will demonstrate how the new generation of full-waveform ALS systems can be used to get a much better classification of solid ground and vegetation cover and consequently DTMs, which can be interpreted archaeologically with much more confidence.