California’s fire department has launched an artificial intelligence (AI) system that examines images from 1,000 cameras across the state to detect potential fires and dispatch firefighting forces to the area. In one example of the ALERTCalifornia system’s capabilities, a camera detected a fire that broke out at 3 a.m. in an isolated area of Cleveland National Park, about 80 kilometers east of San Diego.
With the residents of the area sleeping and the darkness of the night covering the smoke, the fire could easily spread. But AI alerted the fire department, which dispatched 60 firefighters and seven vehicles. The fire was extinguished in 45 minutes.
The system, which was developed by engineers at the University of California, San Diego and was based on technology from California-based DigitalPath, uses images from 1,038 cameras installed across California by utilities and energy companies.
California fire officials hope the platform will serve as an example for similar systems in other states and countries as climate change increases the intensity and frequency of wildfires.
To train the system, firefighter Susan Leininger reviews images in which the AI detected fire and tells the machine whether it was right with a simple yes or no.
The system sometimes gives a false alarm if it sees clouds, dust, or even trucks blowing smoke from the exhaust.
The platform became more reliable with a few hundred iterations of the testing process, the fire department said.
In addition to the AI platform, the service collects a wealth of other data, such as forest cover, mapping the ground under vegetation and fuel accumulation.
Airplanes and drones also collect infrared images that go beyond the capabilities of human vision.
