One, two, tree: how AI helped find millions of trees in the Sahara


Tree mapping helps researchers understand deforestation and climate change, however the technologies used often miss trees that aren’t clustered. Researchers, in collaboration with NASA, used high-resolution satellite images, previously only available to commercial entities, to find a surprisingly large number of trees in the Sahara Desert. Using AI deep learning and one of the world’s most powerful supercomputers at the University of Illinois, they identified more than 1.8 billion trees, manually marking nearly 90,000 so the computer could “learn” which shapes and shadows indicated the presence of trees.

Related Stories