Deep Learning-Based Burned Area Mapping Using Bi-Temporal Siamese Networks and AlphaEarth Foundation Datasets

Seyd Teymoor Seydi · 2025

Read Deep Learning-Based Burned Area Mapping Using Bi-Temporal Siamese Networks and AlphaEarth Foundation Datasets by Seyd Teymoor Seydi (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-07852.txt.

Description

Seyd Teymoor Seydi tackles wildfire damage mapping in a solo 2025 paper on remote sensing. His method pairs satellite imagery from the AlphaEarth foundation dataset with a Siamese U-Net, a neural network that compares images from before and after an event. The imagery combines optical and heat-sensing views with labelled ground truth.

Training uses a United States burn severity dataset, and testing covers seventeen regions across Europe. Fast, accurate maps of burned land help agencies respond to fires, track environmental harm, and judge climate effects. By learning from two time points at once, the network aims to separate real burn scars from ordinary landscape change, giving mapmakers a quicker starting point after each fire season.

Catalog record

CREATORSeyd Teymoor Seydi
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE26 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-07852

Browse the full library · About the Tower of Babel