EarthPV

Free and open mapping of photovoltaics around the Earth!
Including Capacity, Growth and Orientation Systems Above 10 kWp

Documentation  ·  Country atlases  ·  Set up a new country  ·  Experiments  ·  Discord

Many countries do not know how much solar power sits on their roofs, or where. EarthPV answers that from space with free data: it detects PV arrays in Sentinel-2 satellite imagery, turns them into capacity estimates, and hands each detection to OpenStreetMap mappers to verify. What they confirm trains the next model. Built only on open models and global datasets, it works for any country.

Watch We are mapping solar panels in every country, EarthPV explained by Open Energy Transition.

Each country gets an interactive evidence atlas: hand-mapped OpenStreetMap installations plus the model's own detections, per 0.1° cell, with a 90% range on the headline figure. Atlases are tiered by calibration; mapping calibration areas is how an atlas moves up, so maps a calibration area.

At Sentinel-2's 10 m resolution a large array has a shape you can trace and a small one does not, so EarthPV runs two instruments and combines them.

• Segmentation, the fine-tuned TerraMind, outlines arrays above roughly 400 m². Every detection becomes a ranked mapping lead, and it is the only instrument for ground-mounted solar at any size.
• roofclf asks a smaller question of every building: does this roof carry PV? A 100 m² array is a handful of mixed pixels, too few to outline but often enough to classify. It is cross-checked with SPPI, a zero-training spectral index, and calibrated on areas where every installation was mapped by hand.

earthpv atlas combines both with the hand-mapped OpenStreetMap layer into a Best estimate, with overlaps removed and a 90% range attached. A country with neither hand-mapped calibration areas nor a register gets the segmentation half only.

The OpenStreetMap loop. Sentinel-2 is free but coarse; Esri, Bing and Mapbox resolve individual panels but only let a person trace from them in the OpenStreetMap editor. So the model reads only Sentinel-2, people read only the high-resolution layers, and what they map becomes openly licensed training data for the next model.

Keeping it current. An update is a re-run on new imagery, and each round starts from a better map: confirmed leads become verified capacity and training data. See Keeping PV data current.

The same detections also give panel tilt and orientation from Sentinel-2 glint and growth since 2021/22. Everything tried, kept or rejected, with the measurement behind each verdict, is in Experiments.

Measured on 30 exhaustively hand-mapped Pakistani calibration areas (123,898 buildings), at the operating point the published atlas uses:

Most rooftop PV above about 50 m² of panel (roughly 10 kWp) is found, 20 to 50 m² is close to a coin flip, and most smaller systems are missed. Detection is recall-first: about 1 in 10 PV-free buildings is also flagged, so detections are leads for human verification, not a finished inventory. Where household systems are a fraction of a pixel, as in France, the small-rooftop half does not work at all.

The goal is a global atlas assembled country by country, each run and checked by people who know the ground. Every input is global, so the intended shape is a fork per country with the results coming back here. There is no combiner that merges countries into one global surface yet; the shared pipeline, atlas format and data-pack layout are what make that possible later.

gh repo fork open-energy-transition/earthpv --clone --remote
cd earthpv && git switch -c atlas/<country>

pixi run python scripts/new_region.py check --bbox <bbox> --iso3 <ISO3>   # preflight
pixi run python scripts/new_region.py add   --aoi <country> --bbox <bbox> --iso3 <ISO3>
pixi run python scripts/new_region.py plan  --aoi <country>               # your runbook

Every stage is resumable; compose, which builds the imagery, is the long pole at days on a home connection. Package the per-cell capacity table with scripts/build_atlas_data_pack.py and ship it as a GitHub Release asset: commit the manifest, never the pack. One step cannot be automated: after national roofclf scoring, check 20 random cells against high-resolution imagery.

Working with an agent such as Claude Code? CLAUDE.md is the repository's agent brief. Let it run the preflight, the config, the long stages and the draft page, and stop it before anything that needs human eyes: drawing or certifying calibration areas, and signing off validation. Full runbook: Contribute your country atlas back.

EarthPV is the software half of TraceTheSun, a pilot run by Open Energy Transition to make PV mapping cost-effective, verifiable, community-driven and local, worldwide, conceived by Tobias Augspurger and Muhammad Awais. The Pakistani results rest on a student team at the Centre for Water Informatics and Technology (WIT), Lahore University of Management Sciences: Laeeba Hafeez Malik, Tayyiba Shafiq, Nimra Aamir Ali and Vania Malik, who co-designed the pipeline and did the mapping, validation and ground-truth work that makes the numbers checkable. Join the community on Discord.

Code MIT. Imagery from Copernicus Sentinel-2; building footprints from VIDA Open Buildings and Overture Maps; labels from OpenStreetMap contributors under ODbL; administrative boundaries from geoBoundaries under CC-BY; solar-panel import figures from Ember under CC-BY-4.0.

Published data outputs (the atlases, capacity tables, raw detections and any other derived dataset offered for download) are derivative databases of OpenStreetMap's solar labels and, via VIDA Open Buildings, of Microsoft and Google building footprints, so under ODbL's share-alike clause they are licensed under the Open Database License (ODbL) v1.0. Any use must attribute © OpenStreetMap contributors, VIDA Open Buildings (CC BY 4.0) for the footprints and, for anything derived from the German validation, the Marktstammdatenregister (Bundesnetzagentur, Datenlizenz Deutschland - Namensnennung - Version 2.0).