Separating natural forests from other tree cover with AI for deforestation-free supply chains
Google Research
Google DeepMind released Natural Forests of the World 2020, an AI-generated map that distinguishes natural forests from plantations and other tree cover using satellite imagery. The model achieved 92.2% accuracy at 10-meter resolution across all land on Earth, trained on over 1.2 million patch locations and validated against an independent global dataset. The map provides a baseline for compliance with deforestation regulations like the EU's EUDR and supports corporate supply chain monitoring, government oversight, and conservation efforts.
Why it matters
Climate & Sustainability
