Artificial intelligence has learned to map out billions of non-forest trees that were not previously seen. The current method of mapping trees would take months or years to get similar results. NASA has created a way to study broader areas and even calculate carbon storage within that area.
One of the fastest supercomputers and advanced machine learning algorithms mapped more than 1.8 billion trees across an area of more than 500,000 square miles. NASA manually marked nearly 90,000 individual trees across various terrain. It then "learned" which shapes and shadows symbolized trees.
"Dry areas are not well mapped, in terms of what density of trees and carbon is there. This is because normal satellites don't see the trees – they see a forest, but if the tree is isolated, they can't see it," says Martin Brandt.
AI learning to calculate carbon storage will support conservation efforts, provide information on rainfall, and help researchers protect the planet.