Meta· Updated
Meta AI and World Resources Institute launched CHMv2, an open source model that maps global forest canopy heights using DINOv3. Accuracy jumped from R² 0.53 to 0.86, sharpening canopy detail and reducing bias for tall trees.
DINOv2 backbone for DINOv3 Sat-L, pre-trained on SAT-493M — a large satellite imagery dataset — alongside an expanded lidar training set and a specialized loss function. The result: R² accuracy jumps from 0.53 to 0.86, with sharper maps and reduced bias for tall trees.The Canopy Height Maps program already supports climate and conservation efforts across the public sector. The UK's Forest Research agency, the EU Joint Research Centre (for the 3 Billion Trees Initiative), and US city planning tools all use the maps — and plan to adopt CHMv2 for more reliable carbon estimates, deforestation tracking, and reforestation verification.
Download CHMv2 on GitHub or explore the global maps on Google Earth Engine.
We’re announcing Canopy Height Maps v2 (CHMv2), an open source model for high-resolution global forest canopy mapping, developed in partnership with the @WorldResources. CHMv2 leverages our DINOv3 Sat-L vision model, specifically optimized for satellite imagery, to deliver substantial improvements in accuracy, detail, and global consistency. 🔗 Learn more: https://t.co/qhpRUA2iBR
Every HeadsUpAI update is written based on its original source and reviewed before it's published. Read our editorial standards →
More like this
MetaMar 28
NVIDIAJun 1
CohereMar 28
Amazon Web ServicesApr 8