Given the abundance of unlabeled Satellite Image Time Series (SITS) and the scarcity of labeled data, contrastive self-supervised pretraining emerges as a natural tool to leverage this vast quantity of unlabeled data. However, designing effective data augmentations for contrastive learning remains challenging for time series. We introduce a novel resampling-based augmentation strategy that generates positive pairs by upsampling time series and extracting disjoint subsequences while preserving temporal coverage. We validate our approach on multiple agricultural classification benchmarks using Sentinel-2 imagery, showing that it outperforms common alternatives such as jittering, resizing, and masking. Further, we achieve state-of-the-art performance on the S2-Agri100 dataset without employing spatial information or temporal encodings, surpassing more complex masked-based SSL frameworks. Our method offers a simple, yet effective, contrastive learning augmentation for remote sensing time series.
@article{arxiv.2506.18587,
title = {Resampling Augmentation for Time Series Contrastive Learning: Application to Remote Sensing},
author = {Antoine Saget and Baptiste Lafabregue and Antoine Cornuéjols and Pierre Gançarski},
journal= {arXiv preprint arXiv:2506.18587},
year = {2025}
}
Comments
10 pages, 2 figures, accepted at 42nd International Conference on Machine Learning (ICML 2025) Terrabytes workshop