Semantic Neural Radiance Fields for Multi-Date Satellite Data
Valentin Wagner, Sebastian Bullinger, Christoph Bodensteiner, Michael Arens · 2025
Read Semantic Neural Radiance Fields for Multi-Date Satellite Data by Valentin Wagner, Sebastian Bullinger, Christoph Bodensteiner, Michael Arens (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-2502-16992.txt.
Description
Satellite photos of the same ground on different days disagree: cars move, light shifts, seasons turn. Valentin Wagner, Sebastian Bullinger, Christoph Bodensteiner, and Michael Arens taught a neural scene model in 2025. It builds one three-dimensional map with meaning attached, from many dated images plus rough labels for each pixel. The model learns the scene's geometry and what each part is at the same time.
The meaning layer helps the model clean up noisy input labels, and it steadies color prediction where passing objects would otherwise smear the scene. The team also release a hand-labeled dataset for common multi-view satellite scenes, with code, to push the area forward.
Catalog record
| CREATOR | Valentin Wagner, Sebastian Bullinger, Christoph Bodensteiner, Michael Arens |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 32 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2502-16992