WildSAT: Learning Satellite Image Representations from Wildlife Observations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Daroya, Rangel, Cole, Elijah, Mac Aodha, Oisin, Van Horn, Grant, Maji, Subhransu
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913979482243072
author Daroya, Rangel
Cole, Elijah
Mac Aodha, Oisin
Van Horn, Grant
Maji, Subhransu
author_facet Daroya, Rangel
Cole, Elijah
Mac Aodha, Oisin
Van Horn, Grant
Maji, Subhransu
contents Species distributions encode valuable ecological and environmental information, yet their potential for guiding representation learning in remote sensing remains underexplored. We introduce WildSAT, which pairs satellite images with millions of geo-tagged wildlife observations readily-available on citizen science platforms. WildSAT employs a contrastive learning approach that jointly leverages satellite images, species occurrence maps, and textual habitat descriptions to train or fine-tune models. This approach significantly improves performance on diverse satellite image recognition tasks, outperforming both ImageNet-pretrained models and satellite-specific baselines. Additionally, by aligning visual and textual information, WildSAT enables zero-shot retrieval, allowing users to search geographic locations based on textual descriptions. WildSAT surpasses recent cross-modal learning methods, including approaches that align satellite images with ground imagery or wildlife photos, demonstrating the advantages of our approach. Finally, we analyze the impact of key design choices and highlight the broad applicability of WildSAT to remote sensing and biodiversity monitoring.
format Preprint
id arxiv_https___arxiv_org_abs_2412_14428
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle WildSAT: Learning Satellite Image Representations from Wildlife Observations
Daroya, Rangel
Cole, Elijah
Mac Aodha, Oisin
Van Horn, Grant
Maji, Subhransu
Computer Vision and Pattern Recognition
Machine Learning
Quantitative Methods
Species distributions encode valuable ecological and environmental information, yet their potential for guiding representation learning in remote sensing remains underexplored. We introduce WildSAT, which pairs satellite images with millions of geo-tagged wildlife observations readily-available on citizen science platforms. WildSAT employs a contrastive learning approach that jointly leverages satellite images, species occurrence maps, and textual habitat descriptions to train or fine-tune models. This approach significantly improves performance on diverse satellite image recognition tasks, outperforming both ImageNet-pretrained models and satellite-specific baselines. Additionally, by aligning visual and textual information, WildSAT enables zero-shot retrieval, allowing users to search geographic locations based on textual descriptions. WildSAT surpasses recent cross-modal learning methods, including approaches that align satellite images with ground imagery or wildlife photos, demonstrating the advantages of our approach. Finally, we analyze the impact of key design choices and highlight the broad applicability of WildSAT to remote sensing and biodiversity monitoring.
title WildSAT: Learning Satellite Image Representations from Wildlife Observations
topic Computer Vision and Pattern Recognition
Machine Learning
Quantitative Methods
url https://arxiv.org/abs/2412.14428