Скачать или смотреть Global Tree Height Map using DiNOv2 AI, Carbon MRV, Restoration Data by World Resources Institute

World Resources Institute’s Land & Carbon Lab, in collaboration with Meta’s AI Research team, has produced the first global tree-canopy height map at half-meter resolution. The project draws on a foundation model trained on eighteen million high-resolution satellite images, allowing depth estimation from raw RGB data to infer vegetation stature across every continent. This openly licensed product underpins a new phase of spatially explicit forest intelligence and is already being showcased in demonstrations at AI for Good 2025 #aiforgood in Geneva. More information at https://www.wri.org/initiatives/land-...

In technical terms the mapping pipeline builds on the self-supervised DiNOv2 vision backbone, fine-tuned for dense depth prediction and validated against NEON LiDAR ground truth. The model, deployed on petascale compute, achieves a mean absolute error of about 2.8 m and processed the entire landmass within days. The resulting 2020 baseline reveals that canopy taller than one metre occupies roughly fifty million square kilometres, with eighty percent of source imagery captured between 2018 and 2020. Data and weights are mirrored on AWS Open Data and Google Earth Engine, making them immediately callable from common geospatial libraries for further research

Because height is a reliable proxy for above-ground biomass, the half-meter grid supports rapid monitoring, reporting and verification for restoration projects, voluntary carbon markets and corporate net-zero claims. Practitioners can confirm sapling survival within the first two growing seasons, quantify biomass accumulation without field crews and transparently communicate progress through shareable dashboards, providing consistent evidence for funders and regulators worldwide

The team plans to rerun the entire workflow with the forthcoming DiNOv3 release, which promises better stitching of image footprints, recovery of saturated returns in extremely tall forests and gap-filling under persistent cloud. Annual refreshes would transform the static 2020 layer into a living time-series, giving communities, investors and policymakers a near-real-time feedback loop on nature-based climate initiatives. Hosting on Earth Engine and S3 ensures that each update slots directly into existing analytics pipelines, sustaining a resilient open-source monitoring pipeline.

https://sustainability.atmeta.com/blo...

My full AI For Good 2025 video coverage is here:    • AI For Good 2025 video coverage  

This video was filmed using the DJI Pocket 3 ($669 at https://amzn.to/4aMpKIC using the dual wireless DJI Mic 2 microphones with the DJI lapel microphone https://amzn.to/3XIj3l8 ), watch all my DJI Pocket 3 videos here    • Filmed with DJI Pocket 3  

Join https://www.youtube.com/charbax/join for Early Access to my videos and to support my work, or you can Click the "Super Thanks" button below the video to send a highlighted comment under the video! Brands I film are welcome to support my work in this way 😁

Check out my video with Daylight Computer about their revolutionary Sunlight Readable Transflective LCD Display for Healthy Learning:    • Daylight Computer CEO interview: Healthy C...