SAR4Wildfire: Sentinel-1 SAR Time Series for Near Real-Time Wildfire Monitoring with Deep Learning
The objective of this research is to develop a novel, automatic and effective method, using Sentinel-1 SAR time series and a deep learning framework, for near real-time wildfire progression monitoring and burn severity mapping in the selected 2018 wildfire sites in Sweden and British C olumbia, C anada.
Research shows that human-induced climate changes have led to hot, dry conditions that boost the increase in fire activity in some areas. In both 2017 and 2018, the world witnessed many devastating wildfires across the globe. Hotter and drier summers across North America and northern Europe have resulted in increased wildfire activities in cooler and wetter regions such as Sweden, even north of the Arctic C ircle. Wildfires kill and displace people, damage property and infrastructure, burn vegetation, and cost billions of Euros to fight. Therefore, timely and reliable information on active fire front, fire progression, and damage mapping is critical for effective emergency response and management through night, cloud and smoke.
In collaborations with Swedish Civil Contingencies Agency, Swedish Forest Agency, British Columbia Wildfire Service and Ministry of Forests, Lands, Natural Resource, and Rural Development, the objective of this research is to develop a novel, automatic and effective method, using Sentinel-1 SAR time series and a deep learning framework, for near real-time wildfire progression monitoring and burn severity mapping in the selected 2018 wildfire sites in Sweden and British Columbia, Canada. This research is expected to make a significant contribution toimprove the speed and efficiency of wildfire emergency response, decision making, post-fire forest management and proactive response to climate change and managing natural hazards using new technologies.
Funding Agency: Formas
Period: 2020-2024