Collaborative Research: Research Infrastructure: HNDS-I: Building AI-Based Tools for Continental-Scale Archaeological Surveys

The archaeological record of human history is vast, spanning most of the planet's land masses, and mostly unrecorded, with knowledge derived from a tiny fraction of sites around the world. The lack of data from larger connected regions makes it difficult to understand how fieldwork-derived data from small sites fit within the bigger picture of the human past. This project pursues this big picture through a combination of fieldwork and artificial intelligence to produce an archaeological survey that uses high resolution multi-spectral satellite imagery. The datasets produced by this survey enable research on past human adaptation and social networks on a continental scale. Mapping how human populations and settlements are distributed within geographic regions is a critical step for understanding how societies change and adapt to their surroundings. Archaeology is often the only source of information about human settlement patterns before the Early Modern era, but it is extremely challenging to use the field’s traditional methods to map sites across regions. This project meets that challenge by developing artificial intelligence models to identify archaeological features in high resolution satellite imagery, over an area of nearly two million square kilometers. The project develops new deep learning models that are tuned for feature detection and deployed to identify abandoned structures across vast areas. These models are important for research in diverse fields such as earth and environmental sciences, infrastructure planning, and emergency response. The models' results are combined and audited with observational data from fieldwork in different regions. The models and data are open-source and available to the research community to study long-term trends in human adaptation, settlement, and demography. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.