US State Utilizes AI for Early Wildfire Detection
Rising temperatures and climate change increase the risk of unexpected wildfires.
In response to this natural disaster, a program has been launched called “Alert California AI program”, in collaboration between the California Department of Forestry and Fire Protection (Cal Fire) and University of California San Diego.
Using advanced AI technology, this program identifies unusual patterns from 1,032 rotating cameras to spot potential wildfires. Once detected, emergency services are notified to assess the situation.
The AI system is created by UCSD engineers with AI from DigitalPath. The project initial investment was $20 million received from Cal Fire, with an additional $3,516,000 in the near future.
In July, the program successfully detected a fire starting at 3 am in the remote area of San Diego, Cleveland National Forest east, allowing firefighters to respond swiftly and extinguish the flames within 45 minutes.
The technology utilizes LiDAR scans from aircraft and drones to generate 3D data about surfaces. By combining this with information about tree species, the program learns about forest biomass and carbon content. The Machine Learning model can differentiate between smoke and aerosols using massive amounts of camera data.
This program is also capable of measuring various environmental factors, like in winter it measures atmospheric rivers, burn scars, snowpack, sediment, erosion, water quality, and soil quality.
With the increasing impact of climate change leading to more frequent and intense fires, the Alert California AI program is applicable globally and can address a pressing issue.
The intelligence specialist at Cal Fire, Suzann Leininger, assists the AI in learning by reviewing camera footage and confirming fire detections.
Although the AI model effectiveness is being evaluated, it offers valuable data for private companies and researchers to model fire behavior and study the environment.