How AI is being used to fight wildfires in California
The initiative is a partnership between UC San Diego’s ALERTCalifornia public safety program and the California Department of Forestry and Fire Protection.
A network of over 1,000 cameras equipped with artificial intelligence is being used to monitor the California landscape for wildfires.
The initiative is a partnership between UC San Diego’s ALERTCalifornia public safety program and the California Department of Forestry and Fire Protection, or CAL FIRE.
The system operates by having the camera network monitor the wildland 24/7. It utilizes near-infrared night vision and makes 360-degree sweeps about every two minutes. The feeds are then analyzed by AI to spot smoke and alert fire officials.
Currently, six of CalFire’s 21 units are receiving AI-generated alerts.
THE FUTURE OF WILDFIRE DETECTION IS HERE: ARTIFICIAL INTELLIGENCE
Footage from the camera feeds also allows firefighters to assess the topography around a fire. This gives them an opportunity to see which direction the fire might be burning down. Plus, it allows firefighters to monitor other natural disasters, such as mudslides and flooding.
"What I hope that we can use the technology for is to make it safer for the firefighters to do what they need to do," CAL FIRE Intelligence Specialist Suzann Leininger said to FOX Weather correspondent Max Gorden. "That's really the bottom line is all of this technology is about making firefighters safer in the field."
While the AI system has been beneficial for firefighters, it is not a perfect system. For example, the AI will sometimes flag dust or clouds to fire officials. Because of this, some people are tasked with analyzing the AI’s work and improving the system.
The general public take a look at the camera feeds for themselves at the ALERT California website.
The tactic of AI to fight wildfires is catching on.
A system called FireScout has been implemented throughout the western U.S., along with Australia. Much like the UC San Diego-CAL FIRE system, FireScout utilizes over 1,000 cameras scattered across the landscape to monitor for wildfires.
It then uses AI that has been trained with over 10 million images of real wildfires dating back to 2013 to detect smoke plumes from wildfires during the day and detect flames at night. According to FireScout President Bow Rogers, FireScout can identify a wildfire with 99% accuracy.
As of Tuesday, an estimated 30,466 wildfires have burned 1,174,562 acres in the U.S. this year, according to the National Interagency Fire Center. Causes vary, ranging from lightning, which ignited about 2,500 wildfires, and humans, which ignited about 27,850 wildfires.
The NIFC noted that the states with the most wildfires caused by humans are California, North Carolina, Texas, Florida and Pennsylvania.