Published June 4, 2026
| Version v1
Book chapter
Open
Understanding Active Fire Detection Uncertainty with Bayesian Neural Networks
Authors/Creators
- 1. Great Lakes Forestry Centre; Natural Resources Canada
- 2. Northern Forestry Centre; Natural Resources Canada
- 3. Earth Observation for Ecosystem Management; Technical University of Munich
Description
This chapter covers active fire detection using Bayesian Neural Networks (BNN), quantifying prediction uncertainty in wildfire mapping from satellite-derived thermal and reflectance data. It presents two case studies and a framework for uncertainty-aware ecological monitoring.
Part of the EarthRISE Applied Artificial Intelligence and Deep Learning Book, Chapter 6: Ecological Process Simulation.
Files
Applied-Artificial-Intelligence-and-Deep-Learning-Book_Understanding active fire detection uncertainty with Bayesian Neural Networks.pdf
Files
(309.3 kB)
| Name | Size | Download all |
|---|---|---|
|
md5:6635d432169c5c31e94ecbc3af133a1b
|
182.9 kB | Preview Download |
|
md5:fc38f911296c5d6e9dcebf3773a33582
|
126.4 kB | Preview Download |
Additional details
Related works
- Is part of
- Book: 10.5281/zenodo.20547797 (DOI)
- Is supplemented by
- Software: https://github.com/NASA-EarthRISE/EarthRISE-Applied-Artificial-Intelligence-and-Deep-Learning-Book/tree/main/06_Eco_Process_Sim/01__Active_Fire_Detection (URL)