A YOLO-Based Workflow for Detecting and Mapping Archaeological Stone Cairns in Satellite Imagery: A Case Study from Western Ennedi, Chad

A deep learning framework using the YOLOv8 model achieves highly accurate detection of archaeological stone cairns in high-resolution satellite imagery, with precision and recall rates exceeding 83.5%. This non-invasive approach offers a scalable solution for documenting vulnerable heritage landscapes, such as those in the Ennedi Massif, where traditional field surveys are impractical due to remote terrain. The model, trained on GPS-verified cairns, demonstrated stable performance across diverse illumination conditions and outperformed YOLOv5n in localization accuracy while maintaining competitive precision and F1-scores. This workflow enables large-scale archaeological prospection and heritage monitoring, providing a critical tool for preserving and studying remote desert environments.

Ghaderpour, E., Nekoulnang, C. D., Noudjiko, H. M., Rossi, P. P., Rotunno, R., & di Lernia, S. (2026). A YOLO-Based Workflow for Detecting and Mapping Archaeological Stone Cairns in Satellite Imagery: A Case Study from Western Ennedi, Chad. Heritage, 9(6), 237. https://doi.org/10.3390/heritage9060237