close
close

AI technology aims to improve wildfire alert times

By Nikki Davidson
Government technology

Human perception has its limits given the unpredictability of a forest fire outbreak. But with artificial intelligence playing a key role in the response, the question arises: Could AI outperform humans in determining how best to deal with these dangerous situations?

The matter is rhetorical for now, as AI is still very new to wildland firefighting. Historically, however, the potential for human missteps is well known.

HUMAN LIMITATIONS IN FIRE RESPONSE

It can be challenging for people to make informed decisions when planning wildfire response.

In 2007, dispatchers initially dismissed 911 calls about smoke near Lake Tahoe, mistakenly believing it was a controlled fire. In reality, it was a wildfire that ultimately destroyed 254 homes and burned 3,100 acres of mountain wilderness.

Even if emergency responders can accurately assess the fire risk, there is no guarantee that help will actually be available.

The National Interagency Coordination Center (NICC), a central resource management resource management center for wildfires and other major disasters across the country, allocates and mobilizes critical resources such as firefighters, aircraft, equipment, and specialized teams for wildfires and other major disasters across the country.

There has been a constant problem over the years: there were not enough crews to keep up with requests. Almost three in 10 requests the NICC received in 2023 “could not be fulfilled”. Due to the lack of available teams, it is critical that resources are directed where they will have the greatest impact.

Additionally, when firefighting crews are available, the human body can only handle so much in a terrain that is more expansive and rugged than a structural fire.

The U.S. Fire Department tracks and collects information about the causes of firefighter deaths in the line of duty in the United States. An analysis of the data by Government Technology found that at least 346 firefighters and crew members have died while fighting forest, bush and grass fires since 1990.

The most common causes of death during wildfire operations were not burns or smoke inhalation, but rather heart attacks caused by stress and overexertion.

Many others died trying to scout the location and conditions of the fire in rugged terrain, such as Curtis Jessen, a district forester in North Carolina who slipped and fell off a cliff while trying to find out if it was in There was no longer a threat of bushfire in the area.

COULD AI PROVIDE A SOLUTION?

Pano AI, a company focused on wildfire detection and containment, believes technology can revolutionize the way wildfires are fought. It uses deep learning to detect smoke from new fires and powerful cameras to provide powerful video insights that enable authorities to create smarter response plans.

Pano AI has customers in nine states, including city, county, state and federal governments as well as private companies.

The product consists of a network of ultra-high-definition cameras stationed at high altitude and constantly rotating to provide a 360-degree view, coupled with high-speed wireless connectivity, satellite transmissions and cloud-based software.

“We use these feeds to detect the early signs of a new fire, day or night, and then issue alerts that include critical situational awareness video feeds that allow first responders to make early decisions about the fire,” Pano said AI CEO Sonia Kastner.

The San Francisco-based company has served government clients across the United States for three full wildfire seasons, and one of its first was the Aspen Fire Protection District in Colorado in 2021.

Fire’s challenge to AI is complex, Kastner said, but with each season the technology gets smarter.

“Actually, detecting fires is a pretty difficult machine learning problem to solve because wildfires don’t happen very often. “In fact, there are many phenomena in nature that resemble fire, and forest fires look a little different every time,” she said. “The model is getting better and better.”

WILL AI CHANGE WHAT PEOPLE DO?

Pano AI has a 24/7 Threat Intelligence Center, a team of human analysts that review alerts and time-lapse footage and filter out false alarms like clouds before they alert their customers. Using the powerful cameras, analysts differentiate between smoke that may not need immediate attention and incidents that require a multi-agency approach from the start.

“It allows the most serious fires to rise above the noise and means that with agencies that have limited resources, the riskiest incidents can be brought under control quickly,” Kastner said.

One of Pano AI’s customers, the Washington State Department of Natural Resources, told Government Technology that the technology complements people’s work and allows employees to spend more time on other tasks.

“They can focus on the actual dispatching side of the job instead of having to multitask – allowing them to focus on their job as a dispatcher and not so much on gathering information,” said DNR Assistant Division Manager of Plans and Information, Wildland Fire Management Department Angie Lane.

FUTURE OUTLOOK

Pano AI acknowledges that there is a major obstacle to large-scale adoption across the country: budget constraints.

A pilot program at the Telluride Fire Protection District in Colorado last year came with a $100,000 price tag. Pano AI charges its customers about $50,000 per camera per year. Its CEO said funding should come at a higher level – a process that is already being considered.

“City budgets are tight and in principle, in our opinion, this solution should be provided at the state and federal level,” said Kastner. “There has been a lot of discussion and initial efforts in Congress to fund this at the federal level for Forest Management lands, for BLM (Bureau of Land Management) lands. We strongly advocate that funding must continue and be used broadly by federal agencies throughout the Federal Forest Service and BLM lands.”

(c) Government Technology 2024
Visit Government Technology at www.govtech.com
Distributed by Tribune Content Agency, LLC.