NASA Earth Observatory’s map of Earth’s forest canopy height is based on estimates from a deep-learning model applied to Sentinel-2 optical data.
Lang, together with colleagues Konrad Schindler and Jan Wegner, produced the map by merging lidar data from NASA’s Global Ecosystem Dynamics Investigation (GEDI) mission with optical imagery from the European Space Agency’s Sentinel-2 satellites. GEDI’s lidar profiles give detailed canopy heights, but the profiles cover limited areas; Sentinel-2 optical data has abundant coverage, but it is not designed to measure canopy height. The researchers used the GEDI data to train a deep-learning model capable of estimating canopy heights from Sentinel-2 images anywhere on Earth.
Previously: New Map of Global Forest Heights.