How AI-powered cameras are stopping wildfires in Colorado
On a hot, dry July morning, lightning sparked a wildfire in a Colorado canyon 37km south of Denver. Just 24 minutes later, first responders were at the scene dumping water from two helicopters and a 40-person ground crew had mobilised to hike deep into the terrain.
By evening, the fire had been extinguished, after growing to just 12,000 sq km. The risk of a hard-to-control, fast-moving blaze that could engulf nearby homes had been prevented.
Such a destructive prospect is increasingly common in wildfire-prone areas, such as the western regions of the US, the Mediterranean, Australia and Argentina. The reason why Colorado could muster such a speedy response that day? It has two 360-degree cameras mounted on mobile-phone towers. Made by Pano AI, a San Francisco-based start-up that was founded in 2020, the ultra-high-definition cameras scan, identify and monitor wildfires.
In the Mountain state, two electricity utility companes had separately invested in the Pano system, which costs $50,000 (€45,000) a year per camera. The two cameras triangulated the fire’s co-ordinates and a staffer at Pano’s intelligence centre verified the image’s accuracy to confirm that the cameras had not misidentified fog or clouds as smoke.
“We combine human and artificial intelligence to deliver the best quality of alerts to our customers,” says the company’s co-founder and CEO, Sonia Kastner. Historically, wildfire detection has relied on trained spotters in lookout towers deep in the mountains. Often, there isn’t any available land to put up new fire towers in areas near urban populations. “However, there are plenty of telecommunication towers, water tanks and other existing structures that we can mount cameras on,” says Kastner.
With $45m (€40.8m) in venture capital funding to date, Pano has sold its kit to 10 US states, from California to Arizona; it has sold five to Australia and one to a Canadian province. The company’s 30 or so customers include timber companies, ski resorts, government land agencies, energy companies, wind-farm operators and private landowners.
As wildfires become a year-round threat, technology such as Pano’s has the potential to play a crucial role. Demand for such systems, for which existing infrastructure is used and AI complements rather than replaces human know-how, is only likely to grow.
