This flying fire sensor could help track wildfires from a satellite in space

As wildfires currently devastate western North America, a new airborne project team hopes to develop a space solution to stop conflagrations before they get out of control. The project could one day help future firefighters acquire "fire behavior" maps within 20 minutes of an outbreak, using satellite data combined with machine learning (a kind of artificial intelligence), according to a statement from the University of California, Berkeley. The project, funded by a $1.5 million grant, will fund "spotter planes" with infrared detectors -- heat-seeking sensors to examine flame length and geometry to learn more about how fires spread. Meanwhile, machine learning algorithms -- provided they are trained well on other "hot spot" datasets -- could spot new fires in the region within milliseconds, to send alerts. If all goes well in airborne testing, the detector team -- which includes UC Berkeley's Space Sciences Laboratory and Nevada-based fire assessment company Fireball Information Technologies -- hopes to send similar sensors to space within four years to make monitoring and discovery a 24/7 activity.