Landsat-Bench: Datasets and Benchmarks for Landsat Foundation Models

TL;DR

This paper introduces Landsat-Bench, a set of benchmarks for Landsat imagery, enabling standardized evaluation of foundation models and demonstrating the superiority of SSL4EO-L pretrained models over ImageNet in remote sensing tasks.

Contribution

Landsat-Bench provides the first comprehensive benchmarks for Landsat foundation models, including standardized evaluation methods and baseline results.

Findings

SSL4EO-L pretrained GFMs outperform ImageNet models in downstream tasks.

Baseline models achieve +4% OA and +5.1% mAP improvements.

Landsat-Bench facilitates future research in Landsat-based geospatial modeling.

Abstract

The Landsat program offers over 50 years of globally consistent Earth imagery. However, the lack of benchmarks for this data constrains progress towards Landsat-based Geospatial Foundation Models (GFM). In this paper, we introduce Landsat-Bench, a suite of three benchmarks with Landsat imagery that adapt from existing remote sensing datasets -- EuroSAT-L, BigEarthNet-L, and LC100-L. We establish baseline and standardized evaluation methods across both common architectures and Landsat foundation models pretrained on the SSL4EO-L dataset. Notably, we provide evidence that SSL4EO-L pretrained GFMs extract better representations for downstream tasks in comparison to ImageNet, including performance gains of +4% OA and +5.1% mAP on EuroSAT-L and BigEarthNet-L.