SSL4EO-L
Provides pretrained encoders for Landsat imagery back to Landsat 4 and 5.
- Field
- Remote sensing
- Method
- Foundation model
- Data it takes
- imagery
- License
- MIT code, CC0-1.0 weights
SSL4EO-L is a set of self-supervised image encoders for Landsat, with separate weights for Landsat 4–5 TM (top-of-atmosphere reflectance) and for Landsat 7 ETM+ and Landsat 8–9 OLI/TIRS (top-of-atmosphere or surface reflectance). They produce embeddings, or serve as a frozen backbone for segmentation and classification.
The TM and ETM+ weights reach back before Landsat 8, so work spanning decades can use encoders trained the same way on each sensor.
The code is MIT through TorchGeo, and the weights are CC0-1.0.
Catalog entry last checked 2026-10-03. All models →
Run SSL4EO-L on your data
The lab runs this on its own compute, in a container, with the inputs and parameters recorded alongside the result. Initial scoping conversations are free.
Contact the Lab