Technology company pitches AI software pilot program to monitor wildfires in the RDEK
The Regional District of East Kootenay is mulling a proposal from a Canadian tech company that recently pitched it’s artificial intelligence (AI) software and associated hardware for early wildfire detection and response in the region.
Firemark AI recently presented a pilot project to RDEK baord directors that would include a network of cameras and sensors designed to monitor areas outside communities for wildfires while linked up with a software network that incorporates real-time datasets.
The pilot project would run for one year, starting in September and going until the fall of 2025.
“What we want to provide as part of our pilot to the valley is to provide software and offer predictive risk mapping,” said Zubin Kothawala, CEO of Firemark AI, in a presentation to the RDEK board on Aug. 9.
“We are going to cater a software to the region and what we want to do is have input from your first responders and have meetings every month to understand their requirements so our engineers can adapt the software to your needs.”
Adam Preece, with the Firemark AI delegation, spoke about the technological capabilities of the long-range cameras and sensors, noting two cameras at a pilot site up near Fort McMurray autonomously picked up on lightning strikes and started monitoring the area. The sensor package can detect smokeless fires, has a range of 3.5 kms, has satellite and cellular connectivity, is highly portable and can be fully integrated into a user platform.
The cameras were designed and built by a Cranbrook company, Infiniti Electro-optics, a division of Ascendant Technology Group.
Further capabilities could include drone surveillance and mapping at night.
“Our cameras and all our sensors actually use this real-time information and start to change their behaviours,” said Preece. “For example, all of the cameras will start watching those lightning strikes, not for just one day but for multiple days because we know there’s more likelihood that a fire would be there. We integrated satellite imagery from multiple sources so we can use that as detection as well as perimeter mapping and keep real time reports coming to the people who need that information.
“We built our platform to be really agnostic, which enabled us to take advantage of the fire sciences market, where we’re seeing a lot of new sensors, smoke detectors and sensor technology that we can integrate seamlessly into our platform.”
From a financial perspective, the pilot program would cost $156,000 for the year, which would include two sites with cameras, sensors, and access to the software license. Additional pilot sites would cost $66,000 annually, while larger communities such as Cranbrook and Invermere would require more equipment that could be negotiated.
The pilot would be a partnership between Firemark AI and Ashley O’Neil, with ʔa•kinq’uku Fire Service based in ʔaq̓am.
While RDEK board directors didn’t make an immediate decision, the proposal was sent to staff for review and input.
As is usually the case, the push and pull of government jurisdiction, responsibility and funding was a talking point around the board table.
Some wanted the proposal to be pitched to the B.C. Wildfire Service, while others were content to send it to RDEK staff for analysis before making a financial commitment.
Kimberley Mayor Don McCormick said recent severe wildfire seasons are causing high anxiety amongst East Kootenay residents.
“I think we need to take a look at what the requirement is for our residents and be responsive to what our immediate needs are, because lets face it, if we wait for the provincial government on things like this that aren’t even on their radar, it’s never going to happen,” McCormick said. “We see that in so many other areas.”
“I think the bottom line here is that if we don’t do it, it’s not going to get done.”
Cranbrook Mayor Wayne Price, who spent a career in the fire service before running for office, acknowledged the proposal’s innovation, but noted the overall jurisdiction of the B.C. Wildfire Service’s role in provincial wildfire response and management.
“Wildfire is a provincial responsibility, and I appreciate there are gaps, but it’s [wildfires] going to be out in the rural areas on Crown land,” Price said. “I don’t see the benefit of having a number of contractors all over the province. Get one system in place that works, one point of contact. I think that’s the answer, is get one system in place and lets go with that.”
Kothawala said the pilot proposal was specifically built on a micro-scale for the community level to detect small fires, while adding that the company is also planning on engineering a macro scale proposal for the B.C. Wildfire Service.
The B.C. Wildfire Service says it is exploring the use of AI technology in multiple program areas, noting a Request for Proposal (RFP) for BC Wildfire Service Decision Support Technology just closed on August 13.
“We are open to exploring the use of AI technology in multiple program areas,” said the B.C. Wildfire Service, in a statement. “However, it will take time to find the technologies that provide value to the organization and the public. There are a multitude of vendors marketing AI services to fire agencies and we have a duty to ensure that the technology is beneficial, trustworthy and cost effective.”
The BCWS is piloting two automated fire growth models, FireCast and Wildfire Analyst, which are are fire growth model applications, and not AI.
FireCast is based on the Canadian Fire Behaviour Prediction System and produces automated fire growth models for all active wildfires using publicly available numerical weather forecasts, fuel and topography layers, and fuel moisture codes provided by BCWS.
Wildfire Analyst uses a US system to account for one-, 10-, and 100-hour fuel moisture conditions, modified Rothermel fuel models, and a number of proprietary data layers as inputs. It produces automated fire growth models for new fire starts.
However, the BCWS Predictive Services does use AI algorithms to interpret satellite imagery to determine fire boundaries when it is not practical to use airborne GPS or infrared (IR) scanning to do so. Predictive Services is also exploring the use of AI for enhancing weather forecasts and predicting fire occurrence for preparedness planning.
“This is perhaps the best example of how BC Wildfire Service plans to use AI at a macro scale to monitor fire growth throughout the province,” said the BCWS.
