Detecting Illegal Extraction in the Amazon Rainforest with Geospatial Foundation Models

Tools

Fergus-Allen, Cameron and Werkmeister, Astrid and Macdonald, Malcolm (2026) Detecting Illegal Extraction in the Amazon Rainforest with Geospatial Foundation Models. In: High Performance and Disruptive Computing in Remote Sensing School 2026, 2026-06-09 - 2026-06-12. (Unpublished)

[thumbnail of Detecting-Illegal-Extraction-in-the-Amazon-Rainforest-with-Geospatial-Foundation-Models-HDCRS-2026-Poster]
Preview
 Text. Filename: Detecting-Illegal-Extraction-in-the-Amazon-Rainforest-with-Geospatial-Foundation-Models-HDCRS-2026-Poster.pdf
Final Published Version
License: Strathprints license 1.0
Download (17MB)| Preview

Abstract

Illegal extraction in the Amazon Rainforest, such as logging and mining, devastates natural ecosystems. Their vast geographical footprint is best monitored with Earth Observation (EO) data. Recent advances in geospatial foundation models (GFMs) use self-supervised learning to increase the accuracy of various types of geospatial analysis, especially in low-label conditions. For instance, geospatial embeddings enhance the pixel-wise classification of mining sites. Looking forward, novel physics-aware GFMs will enable the improved detection of illegal activities with fine-grained footprints, like selective logging.

ORCID iDs

Fergus-Allen, Cameron ORCID logoORCID: https://orcid.org/0009-0002-7504-0704, Werkmeister, Astrid ORCID logoORCID: https://orcid.org/0000-0002-0174-5851 and Macdonald, Malcolm ORCID logoORCID: https://orcid.org/0000-0003-4499-4281;
CORE (COnnecting REpositories)