The terrific fires typifying the last few fire seasons in California’s forests have had scientists brainstorming new and better ways to combat them, and one effort from the startup Salo Sciences, called the California Forest Observatory (CFO), uses satellites and AI to create a map covered in small green points.
Each of these points represents a real tree, which together represents one of the most ultra-detailed satellite pictures of the earth’s surface available. This will allow wildfire authorities to predict and plan where fires are starting, where they might be headed, and how damaging they might become using information like wind speed and direction, canopy height, and ground vegetation cover.
The CFO creates these pictures from a string of 100 bread-loaf-sized satellite cameras managed by a private company formed by NASA scientists called Planet.
Spinning in a line around the earth, they act in unison like a barcode scanner of our planet’s surface, taking images at about three meters per pixel, an order of magnitude higher than those taken from larger, older satellites.
The satellites are then aided by lidar, a laser technology that gathers information, mounted on a plane, before both streams of data are fed into a deep-learning algorithm which determines things like tree height.