University of West Florida researchers are now using artificial intelligence to track data that will improve how damage from wildfires is mapped.
UWF Ph.D. student Valeria Martin has introduced what’s being called CalFireSeg-50, a dataset that was formulated from satellite imagery and data from 50 of the largest wildfires in California between 2019 and 2023, said a UWF news release.
Martin conducted her research with assistance from Brent Venable, UWF Director of the Intelligent Systems and Robotics Doctoral Program at UWF, and Derek Morgan, UWF Associate Professor of Earth and Environmental Sciences. They collected satellite images and provided the examples to AI, which developed recognition of fire-damaged areas. Then, using the images, the AI developed mapping systems.
“By pairing satellite data with deep-learning models, we can detect wildfire damage with high accuracy,” Martin said. “The insights from this work can support emergency response, environmental monitoring and long-term recovery planning.”
The AI programming helps develop models to indicate where fires might burn the hottest. That data helps emergency responders understand how wildfires burn and spread while pinpointing areas of priority. The AI imagery and maps also help analyze damage and track vegetation patterns across land after a blaze in the recovery phase.
“This project showcases how advanced GeoAI techniques can meaningfully support environmental monitoring,” Venable said. “Valeria’s exceptional work demonstrates the power of interdisciplinary research and reflects the innovative spirit of the Intelligent Systems and Robotics doctoral program.”
Martin’s research is gaining more attention and observation. She presented the Findings of her research at the 13th annual Association for Computing Machinery Special Interest Group on Spatial Information conference that took place in Minneapolis in November.
