Artificial intelligence stands to change wildfire recovery and containment forever during a time when fires in the West are doing things that seasoned firefighters of 40 years have never seen.
"You could not hire enough firefighters," stated Brent VanKeulen, deputy director of the Western Fire Chiefs Association (WFCA). "You can't fly enough planes. You can't get enough dozers on the ground to meet the challenge of what we're facing now."
Over the last year, public safety teams have tested a new tool in western states like California and Oregon, and results have been promising. The tests involve imagery taken by a 360° camera both before and after a fire. Thanks to cloud computing and machine learning, that visual product can be transformed into mapping data that shows "what was damaged, where it was damaged and how badly it was damaged," said David Blankinship, chief technology adviser with WFCA.
Perhaps the most impressive aspect of the tool is the resources it saves. The machine learning component parses massive amounts of data in a matter of hours — something that could take weeks with "hundreds of inspectors with clipboards or phone apps," Blankinship remarked.
"We're comfortable with saying we can image these fires and turn this data into a useful product within 24 hours," he added.
WFCA wants to push the boundaries to where the product could come in less than 24 hours. In fact, it's possible results could eventually be seen in real time, according to Blankinship. To that end, WFCA is working to find the best hardware sensors, software processing and algorithms to expedite the process.
Although WFCA couldn't reveal the private organisation with which it has collaborated on these tests, one might infer from a recent Techwire article that Google is involved.
The new tool can be a "game-changer" for communities that need to rebuild after fire disasters, said Robert Horton, fire chief for Jackson County Fire District #3 in Oregon. The challenge to recovery lies in assessing what resources, such as the number of dump trucks, need to be brought into a damaged community to start the cleanup and rebuilding process.
Horton remembers the effort it took a traditional task force to assess damages wreaked by the Almeda Fire last year in Jackson County.
"It was easily days, if not a week, for the task force to navigate through all the buildings and provide their damage assessment," Horton recounted. "The team with the pilot technology drove through in maybe 2.5 or three hours."
Horton said the new tool can aid firefighting by freeing up task forces. The logic is simple: Any task forces that would otherwise have to be assigned to a recovery effort can instead be sent to help other teams contain active fires.
In Horton's eyes, this type of tech could play a part in a larger support system when it comes to protecting and assisting community members.
"A lot of these houses that were burned (by Almeda) were owned by low-income people in Jackson County... there was simply nowhere else for these people to go, so they leave our community," Horton explained. "To me that's heartbreaking. They're part of our community... I think we can do better, and we know this technology is a great step in building resiliency in our community to help prevent fires but also help recover from disasters when they do strike."
While innovation in tech might lead to better outcomes against fires, VanKeulen said the real innovation is in the culture shift he has seen in public safety in terms of its motivation to go after the latest solutions. In the last two years, he has observed an engagement with tech that has no precedent. Public safety's willingness to try new things has drawn attention from both tech companies and venture capitalists, creating a "big stir that didn't exist, say, five years ago."
"If you look at public safety and its actual adoption of technology compared to what a regular consumer would consider to be innovative in their phone or wearable, they (consumers) are using really old stuff," VanKeulen concluded. – Governing/Tribune News Service