arXiv:2511.02831v1 Announce Type: new 
Abstract: The number and diversity of remote sensing satellites grows over time, while the vast majority of labeled data comes from older satellites. As the foundation models for Earth observation scale up, the cost of (re-)training to support new satellites grows too, so the generalization capabilities of the models towards new satellites become increasingly important. In this work we introduce GeoCrossBench, an extension of the popular GeoBench benchmark with a new evaluation protocol: it tests the in-distribution performance; generalization to satellites with no band overlap; and generalization to satellites with additional bands with respect to the training set. We also develop a self-supervised extension of ChannelViT, ChiViT, to improve its cross-satellite performance. First, we show that even the best foundation models for remote sensing (DOFA, TerraFM) do not outperform general purpose models like DINOv3 in the in-distribution setting. Second, when generalizing to new satellites with no band overlap, all models suffer 2-4x drop in performance, and ChiViT significantly outperforms the runner-up DINOv3. Third, the performance of all tested models drops on average by 5-25\% when given additional bands during test time. Finally, we show that fine-tuning just the last linear layer of these models using oracle labels from all bands can get relatively consistent performance across all satellites, highlighting that the benchmark is far from being saturated. We publicly release the code and the datasets to encourage the development of more future-proof remote sensing models with stronger cross-satellite generalization.

GeoCrossBench هو معيار جديد مصمم لتعزيز قدرات التعميم لنماذج مراقبة الأرض مع زيادة عدد الأقمار الصناعية للاستشعار عن بعد. يتناول هذا التطور التحديات التي تطرحها التنوع المتزايد للأقمار الصناعية والحاجة إلى تدريب فعال على بيانات جديدة.

GeoCrossBench es un nuevo benchmark diseñado para mejorar las capacidades de generalización de los modelos de observación de la Tierra a medida que aumenta el número de satélites de teledetección. Este desarrollo aborda los desafíos que plantea la creciente diversidad de satélites y la necesidad de un entrenamiento efectivo con nuevos datos.

GeoCrossBench est un nouveau benchmark conçu pour améliorer les capacités de généralisation des modèles d'observation de la Terre alors que le nombre de satellites de télédétection augmente. Ce développement répond aux défis posés par la diversité croissante des satellites et la nécessité d'une formation efficace sur de nouvelles données.

GeoCrossBench is a new benchmark designed to enhance the generalization capabilities of Earth observation models as the number of remote sensing satellites increases. This development addresses the challenges posed by the growing diversity of satellites and the need for effective training on new data.

GeoCrossBench: Cross-Band Generalization for Remote Sensing

Was this article worth reading? Share it

Ready to build your own newsroom?