Cutting Through the Green: A Case for Grassland Archaeology Using UAV Multispectral Data

Abstract

Advances in low-altitude remote sensing are needed to improve the effectiveness of archaeological prospection in the Netherlands. The geomorphological situation and land use history make applying various remote sensing and geophysical technologies particularly challenging. Technologies developed for precision agriculture have improved the field. However, the technology has yet to deliver reliable results when used in archaeological prospection in the Netherlands. The Parrot Sequoia multispectral camera was used to study a known archaeological site near Raalte, the Netherlands, to investigate the particular influence of land use practices on archaeological vegetation mark visibility. The site was beneficial for this purpose because the archaeological structures beneath the surface are well known. Also, the hydrological situation is relatively constant throughout the year. This site, therefore, is the perfect test site because certain variables are known constants. The four missions resulted in 24 ortho-mosaics, of which the RGB, the NDVI and the SR datasets were selected for further analysis. In the RGB datasets, vegetation marks were only visible in the August dataset. Vegetation marks were recognized in the June and August NDVI and SR datasets. All of the detected vegetation marks in the NDVI and SR datasets were visible at times shortly after the grass had been mown, and none were visible when the grass was at the peak of its growth. This study has made a contribution to better understanding the visibility of vegetation marks regarding their timing in the land use cycle of meadows and pastures.