Also something we could have used in Forestry applications.
Neural Network Can Read Tree Heights from Satellite Images
ETH Zurich (Switzerland)
Stéphanie Hegelbach, April 20, 2022
Researchers at Switzerland's ETH Zurich leveraged an artificial neural network and satellite images to develop a high-resolution global vegetation height map for 2020 that could be used for sustainable regional development planning or to assess carbon emissions associated with deforestation. The map allows users to determine tree heights on any piece of woodland on Earth at a resolution of as little as 10x10 meters per pixel. The convolutional neural network was trained using millions of images from the European Space Agency's two Copernicus Sentinel-2 satellites, along with tree height data based on space laser measurements from NASA's Global Ecosystem Dynamics Investigation mission. The map will be made public along with its source code. ...
No comments:
Post a Comment