A group of Cal Poly professors and students have created an artificial intelligence (AI) technology used to assess damaged buildings and homes after wildfires. The team hopes that their technology can be used to swiftly assess damaged buildings and aid thousands of California residents in receiving insurance benefits.
The technology, Damage Map, uses aerial images to scan large areas and create damage reports just minutes after the smoke clears from the fire. This would ideally get information into the hands of fire departments, insurance companies and homeowners immediately. The Damage Map team says their goal is to build something that can be a great service to California, and eventually the world.
According to Cal Fire, 10,488 structures were damaged or destroyed due to California wildfires in 2020. However, for the thousands of people who lost their homes and businesses, financial help and insurance coverage was not received until weeks after. This wait is partially due to the damage assessment process, which requires an individual to physically assess the destruction to each structure before any insurance claim can be filed.
Damage Map is an application that takes an aerial image after a wildfire and automatically locates the buildings, classifying them as either damaged or undamaged. The application can be used by anyone, so long as they have access to a computer, Wi-Fi and an aerial image of their home.
Social sciences professor Andrew Fricker got the idea for Damage Map after his parents were evacuated from their home during “Camp Fire,” California’s most deadly fire to date.
Almost a month had gone by before Fricker, his parents and the surrounding community were informed about the condition of their homes. In a state of panic and concern, Fricker and a childhood friend even considered hiking back into the neighborhood, eventually deciding against it due to safety and legal reasons. Fricker’s family home was still standing, but his friends’ had burned down.
The lack of information for community members, insurance companies and the surrounding fire departments made it difficult for the Frickers to collect insurance payout. According to Fricker, around 30,000 people were in the same position.
“It was really hard, but we were really desperate for information,” Fricker said. “I found myself pouring everything I had into this and trying to use satellite information to figure out what had happened to my home.”
Within months of starting Damage Map, computer science professor Jonathan Ventura, industrial engineering senior Gustave Rousselet and a team of Stanford doctoral students joined Fricker in his search to solve the damage assessment dilemma.
The application showed to be 90% accurate compared to official damage reports when tested on images from California’s Camp Fire and Carr Fire.
Damage Map is less likely to be used by everyday people since it requires aerial footage, a computer and Wi-Fi. However, fire departments have access to these resources, which means they could greatly benefit from Damage Map.
“I think it’s an idea that solves a real problem and it’s awesome that we could apply technology to create such a strong social impact,” Ventura said.
As the threat of wildfires increases, so does the possibility of community members losing their homes, however, the Damage Map team believes that if partnered with the right people, their technology could revolutionize emergency response to wildfires and other natural disasters for good.
“We have proved that it can work and that it can be effective through our prototype, but it needs more development, it needs more training,” Fricker said. “Our hope is that we can partner with people at Cal Poly to try and keep this alive.”