Saved in:
Bibliographic Details
Main Authors: Lahrichi, Saad, Sheng, Zion, Xia, Shufan, Bradbury, Kyle, Malof, Jordan
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2502.10669
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909495676895232
author Lahrichi, Saad
Sheng, Zion
Xia, Shufan
Bradbury, Kyle
Malof, Jordan
author_facet Lahrichi, Saad
Sheng, Zion
Xia, Shufan
Bradbury, Kyle
Malof, Jordan
contents Self-supervised learning (SSL) has demonstrated significant potential in pre-training robust models with limited labeled data, making it particularly valuable for remote sensing (RS) tasks. A common assumption is that pre-training on domain-aligned data provides maximal benefits on downstream tasks, particularly when compared to ImageNet-pretraining (INP). In this work, we investigate this assumption by collecting GeoNet, a large and diverse dataset of global optical Sentinel-2 imagery, and pre-training SwAV and MAE on both GeoNet and ImageNet. Evaluating these models on six downstream tasks in the few-shot setting reveals that SSL pre-training on RS data offers modest performance improvements over INP, and that it remains competitive in multiple scenarios. This indicates that the presumed benefits of SSL pre-training on RS data may be overstated, and the additional costs of data curation and pre-training could be unjustified.
format Preprint
id arxiv_https___arxiv_org_abs_2502_10669
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Is Self-Supervised Pre-training on Satellite Imagery Better than ImageNet? A Systematic Study with Sentinel-2
Lahrichi, Saad
Sheng, Zion
Xia, Shufan
Bradbury, Kyle
Malof, Jordan
Computer Vision and Pattern Recognition
Self-supervised learning (SSL) has demonstrated significant potential in pre-training robust models with limited labeled data, making it particularly valuable for remote sensing (RS) tasks. A common assumption is that pre-training on domain-aligned data provides maximal benefits on downstream tasks, particularly when compared to ImageNet-pretraining (INP). In this work, we investigate this assumption by collecting GeoNet, a large and diverse dataset of global optical Sentinel-2 imagery, and pre-training SwAV and MAE on both GeoNet and ImageNet. Evaluating these models on six downstream tasks in the few-shot setting reveals that SSL pre-training on RS data offers modest performance improvements over INP, and that it remains competitive in multiple scenarios. This indicates that the presumed benefits of SSL pre-training on RS data may be overstated, and the additional costs of data curation and pre-training could be unjustified.
title Is Self-Supervised Pre-training on Satellite Imagery Better than ImageNet? A Systematic Study with Sentinel-2
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2502.10669