GSR4B: Biomass Map Super-Resolution with Sentinel-1/2 Guidance

Are you interested in aboveground biomass estimation? Would you need some global, high-resolution biomass maps? but can only find some low-resolution ones? Well, turns our you can super-resolve existing biomass rasters up to 10m resolution with the help of Sentinel-1/2 imagery! The resulting estimates are more accurate than direct regression solely from Sentinel-1/2 imagery and less prone to over-estimating low values and under-estimation high values. This work was accepted for an oral presentation at the ISPRS Geospatial Week 2025 in Dubai on April 8 at 14:30 (GMT+4), come see us if you are around!

Website: https://ecovision-uzh.github.io/GSR4B/

Paper: https://arxiv.org/abs/2504.01722

Code: https://github.com/kaankaramanofficial/GSR4B 

Team University of Zurich: Kaan Karaman, Yuchang Jiang, Damien Robert (PhD), Vivien Sainte Fare Garnot, Maria João Santos, Jan Dirk Wegner