Get Full Government Meeting Transcripts, Videos, & Alerts Forever!
Get email alerts on the Community Risk Reduction topic
No spam. Unsubscribe anytime.
La Habra Heights council hears AI wildfire-risk pitch; residents press privacy, accuracy and next steps
Summary
A FireScore AI representative demonstrated parcel-level, 3D wildfire-risk maps that combine satellite, NAIP training and drone confirmation; council and residents asked for clearer public outreach, privacy protections and a plan for how flagged parcels would be inspected and acted on.
A representative from FireScore AI told the La Habra Heights City Council on Aug. 11 that the company’s “physical AI” platform can produce parcel-level, three-dimensional fuel-risk maps by fusing Sentinel-2 satellite data, NAIP training imagery and targeted drone flights for confirmation. The presenter said the system can identify vegetation indices, live fuel moisture and even differentiate vegetation species to prioritize inspections and mitigation for the city.
The company highlighted one area on East Skyline Drive as the highest-priority parcel in its model. “Our system looks at current vegetation, not historic. Here, for example, is biomass from last week, not three years ago,” the FireScore AI representative said during the presentation and added that a drone flight can be used to confirm model outputs before any city action.
Fire Marshal Keith Jorgensen said the technology could help the department locate hazards that are inaccessible from the street. “Over half of the city — we are either unable to access the property or, in terms of the vacant parcels, many of them were inaccessible because of unsafe or steep terrain,” Jorgensen said, noting that some parcels flagged by the model had passed earlier street-level inspections.
Residents and several council members pressed the company and staff for specifics about accuracy, privacy and operational use. Stephen Black (Citron Road) said the map provided in the packet lacked parcel-level clarity and urged the city to give residents an easy way to see how their lot is classified. Other commenters asked how often the model is updated, whether the company’s drone flights would be marked as city operations and what would happen if the model flagged a parcel that the city’s inspectors could not safely access.
The FireScore AI representative said the company uses Sentinel-2 satellite bands downsampled via "super-resolution" methods to approximately one-meter resolution, then confirms selected locations with a single drone pass at roughly 300 feet. The presenter also described a privacy process in which homes, pools and people would be blurred in the imagery and said the platform classifies dogs and humans so they are removed from analysis.
Council reactions were mixed but generally favorable to further study. Staff recommended additional public outreach and a workshop so residents could see maps and ask technical questions before the council considered any formal pilot or contract. The mayor and multiple council members supported scheduling a public workshop and including the city’s Emergency Preparedness Committee (EPC) in outreach.
Next steps: staff will explore public-engagement options and the council directed staff to return with a workshop plan and further information about pilot scope, costs and privacy protections.
