Labeled dataset of Sentinel-1 SAR imagery Despeckled with multitemporal fusions
Jean Pierre Díaz-Paz, Ahmed Alejandro Cardona-Mesa, Paula Andrea Muñoz-Uribe, Rubén Darío Vásquez-Salazar, Santiago Zapata-Vargas · 2026
Read Labeled dataset of Sentinel-1 SAR imagery Despeckled with multitemporal fusions by Jean Pierre Díaz-Paz, Ahmed Alejandro Cardona-Mesa, Paula Andrea Muñoz-Uribe, Rubén Darío Vásquez-Salazar, Santiago Zapata-Vargas (2026) 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/PMC13333279.txt.
Description
Jean Pierre Díaz-Paz and co-authors present a labeled dataset in 2026 for cleaning speckle noise from Sentinel-1 radar images. Radar pictures are grainy by nature, and training smart filters requires matched noisy and clean examples, which are hard to build honestly.
The dataset holds 512 by 512 pixel radar tiles in both VV and VH polarizations, gathered worldwide through an automated Python and Google Earth Engine workflow with random sampling and quality filtering. For each spot, cleaner reference images were made by averaging stacks of 5, 10, 15, 20, and 25 passes, letting users study how averaging depth affects speckle removal. Land cover maps from ESA WorldCover v200 and a metadata file ride along, and benchmark scores including ENL, MSE, SSIM, and PSNR are supplied for sample tiles. The set is hosted on Mendeley Data as a reproducible training resource.
Catalog record
| CREATOR | Jean Pierre Díaz-Paz, Ahmed Alejandro Cardona-Mesa, Paula Andrea Muñoz-Uribe, Rubén Darío Vásquez-Salazar, Santiago Zapata-Vargas |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45302 |
| LICENSE | https://creativecommons.org/licenses/by/4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |