SatMAE: Pre-training Transformers for Temporal and Multi-Spectral Satellite Imagery

TL;DR

SatMAE introduces a novel pre-training framework for satellite imagery that leverages temporal and multi-spectral data, significantly improving downstream task performance through masked autoencoding techniques.

Contribution

It develops a Masked Autoencoder-based pre-training method tailored for temporal and multi-spectral satellite imagery, incorporating temporal embeddings and spectral band encodings.

Findings

Up to 7% improvement in supervised benchmark performance.

Up to 14% enhancement in land cover classification accuracy.

Significant gains in semantic segmentation tasks.

Abstract

Unsupervised pre-training methods for large vision models have shown to enhance performance on downstream supervised tasks. Developing similar techniques for satellite imagery presents significant opportunities as unlabelled data is plentiful and the inherent temporal and multi-spectral structure provides avenues to further improve existing pre-training strategies. In this paper, we present SatMAE, a pre-training framework for temporal or multi-spectral satellite imagery based on Masked Autoencoder (MAE). To leverage temporal information, we include a temporal embedding along with independently masking image patches across time. In addition, we demonstrate that encoding multi-spectral data as groups of bands with distinct spectral positional encodings is beneficial. Our approach yields strong improvements over previous state-of-the-art techniques, both in terms of supervised learning performance on benchmark datasets (up to $\uparrow$ 7%), and transfer learning performance on downstream remote sensing tasks, including land cover classification (up to $\uparrow$ 14%) and semantic segmentation. Code and data are available on the project website: https://sustainlab-group.github.io/SatMAE/