Autodetection of ancient Arabian tombs in high-resolution satellite imagery

High circular tombs (HCTs) in southern Arabia provide valuable information for anthropologists who seek fundamental understanding of the transition of ancient peoples from a nomadic pastoral lifestyle, to agro-pastoralism, and eventually to the formation of ancient states. In particular, knowing the geographical distribution of HCTs across the region informs theories on patterns of territoriality and environmental and social factors that are implicated in the emergence of ancient civilizations. The small size of the HCTs, vast search regions, and rugged terrain make mapping them in the field difficult and costly. In this article, a detection algorithm is described and quantitatively evaluated and establishes the feasibility of automatically detecting these tombs in satellite imagery. By narrowing the search to a smaller set of candidate locations, wide area discovery and mapping can be performed much more effectively.