Call for Papers: Future Sight on Past Landscapes – Vision Foundation Models for Archaeological Remote Sensing and Landscape Archaeology (CAA2026, Session 37)

—

by

Alphaeus Lien-Talks
in

We are pleased to announce Session 37: Future Sight on Past Landscapes – Vision Foundation Models for Archaeological Remote Sensing and Landscape Archaeology, which will take place at the Computer Applications and Quantitative Methods in Archaeology (CAA) Conference 2026 in Vienna.

📅 Abstract submission deadline: 26 October 2025
🔗 Submit your abstract
📜 Full list of sessions (see S37)

About the Session

Vision Foundation Models (VFMs) such as SAM, DINOv3, and geospatial-specialised models like Prithvi, AlphaEarth, and DeepAndes are reshaping how archaeological features are detected and interpreted. Their ability to generalise across imagery types and perform zero- or few-shot detection and segmentation opens new opportunities for landscape archaeology and remote sensing, where annotated datasets remain scarce.

This session invites applied and critical contributions that explore how VFMs can be adapted to archaeological contexts, from satellite and UAV imagery to LiDAR data, while reflecting on reproducibility, transparency, and ethical implications.

Suggested Topics

Submissions may address (but are not limited to):

  • Case studies of VFMs applied to archaeological feature detection
  • Comparative evaluations of general-purpose vs. geospatial models
  • Workflow integration in the field (e.g., drones, edge devices, real-time mapping)
  • Benchmarking, reproducible pipelines, and open-source tools
  • Critical perspectives on interpretability, bias, and failure modes
  • Community-driven infrastructures for annotation, collaboration, and benchmarking
  • Ethical considerations, including site exposure, looting risks, and cultural sensitivity

Why Participate?

This session highlights both the promise and the challenges of applying VFMs to archaeology. By participating, you will help define best practices, reproducible pipelines, and ethical frameworks for the responsible use of AI in remote sensing and landscape studies.

The discussion aligns closely with the objectives of the MAIA COST Action, which seeks to promote transparent, critical, and inclusive uses of AI across heritage research.

Session Organisers

  • Sohini Mallick (Independent Researcher)
  • Jürgen Landauer (Landauer AI Research)
  • Agnes Schneider (Leiden University)