Washington state lands officials say artificial intelligence is helping them to spot and track wildfires this summer.
The Department of Natural Resources is five months into a pilot project in which about a dozen stations with rotating cameras have been installed on state land.
“They are in high-priority areas where we’ve experienced a large number of fires,” said Angie Lane, the assistant manager for DNR’s wildfire division. “That, combined with the topography and the fuels that are located at those sites, makes them prime locations for each detection system.”
Lane spoke at a press conference Thursday in Ellensburg.
The cameras are linked to the system created by Pano AI, a San Francisco company that works with fire agencies around the nation. That system is connected to the rest of the wired world by T-Mobile’s 5G network.
“The artificial intelligence is looking for the very first signs of wildfire smoke,” said Kathryn Williams, the business development manager at Pano AI and a former wildland firefighter.
When the cameras spot what could be a fire, the AI data is reviewed by human inspectors at Pano who monitor the system 24/7.
“We also ingest other data feeds, like NOAA (National Oceanic and Atmospheric Administration) satellites and 911, that are coming into our intelligence center, and we’re using our camera to ground-truth that intelligence before alerting DNR,” Williams said.
If a fire is suspected, Pano contacts DNR, which begins the process of dispatching state and local fire crews and resources to the fire, said Lands Commissioner Hilary Franz.
“Not only is it helping us get on those fires quickly and know how big it is and what kind of resources and what the terrain is, our goal is to be able to get this information down to our public communities, our emergency responders, those who are out there trying to inform the public about where the fire location is. Where is it moving? Do you need to evacuate? What level of evacuation?” she said.
DNR plans to install about 10 more camera stations by the time the pilot program ends next summer.
“I remember here, in 2017, where smoke would happen in the air, somebody would call it in. You didn’t know if it was a big fire or where the fire was and we would send out an engine,” Franz said. “They get to it and, ‘yeah, we’ll probably need four more engines and then we’ll contain it.’ As we know, with this wind and dry fuel load, we’re then way behind the ball on the fire.”
Franz said, if AI technology proves valuable in helping crews get to fires earlier, she’ll look to post more camera stations around the state. The pilot project is funded by state money, but she says her agency will look to develop partnerships with other organizations to expand the network to federal, tribal and private lands.
She says cameras are not only valuable in monitoring wildfires, but could also help land management agencies keep better tabs on their prescribed burns.
