UniverSat: Resolution- and Modality-Agnostic Transformers for Earth Observation
This work addresses the challenge of adapting Vision Transformers to diverse Earth Observation data, offering a unified model that reduces the need for sensor-specific architectures.
UniverSat introduces a Vision Transformer backbone with a Universal Patch Encoder that handles arbitrary spatial, spectral, and temporal resolutions across optical and non-optical sensors, enabling a single model trained on heterogeneous multimodal Earth Observation data to achieve strong results on classification and segmentation benchmarks.
Vision Transformers (ViT) dominate computer vision. However, their reliance on rigid patch projectors hinders transfer to Earth Observation (EO), where input modalities, scales, and resolutions vary widely. We introduce UniverSat, a ViT-style backbone built around a Universal Patch Encoder that maps patches from arbitrary spatial, spectral, and temporal resolutions, and from both optical and non-optical sensors, into a shared embedding space with a shared set of weights. This enables training a single model on heterogeneous multimodal corpora via self-supervision, yielding robust, sensor-agnostic spatial features. We validate this approach with strong results across classification and segmentation on standard EO benchmarks from GeoBench, PANGEABench, and SpectralEarth. Our code and models are available at https://github.com/gastruc/UniverSat.