Open access
Oct 2026
Near-Real-Time Wildfire Segmentation and Brightness Temperature Estimation from Geostationary Satellite Imagery Using Deep Learning
A two-step deep learning approach is introduced that statistically infers VIIRS-like sub-pixel fire segmentation and brightness temperature (BT) from GOES imagery and reduces the fire region BT root mean square error.
Ryota Yagi, Mukul Badhan, M. Bavandpour et al.
· Fire · 0 citations