Google Research· ResearchResearch paper

Separating natural forests from other tree cover with AI for deforestation-free supply chains

Google Research and Google DeepMind release Natural Forests of the World 2020, a globally consistent 10-meter resolution map distinguishing natural forests from other tree cover, published in Nature Scientific Data. The map uses a multi-modal temporal-spatial vision transformer (MTSViT) model and achieves 92.2% accuracy. It aims to support deforestation-free regulations like the EUDR.

From the source

we’re releasing Natural Forests of the World 2020, a new map and dataset, published in Nature Scientific Data. This project ... provides a critical baseline for deforestation and degradation monitoring. It's the first globally consistent, 10-meter resolution map that differentiates natural forests from other tree cover and achieves a best-in-class accuracy of 92.2% when validated against a global independent dataset.

research.google