Abstract
Foundation models are rapidly transforming Earth Observation data mining by enabling generalizable and scalable solutions for key tasks such as scene classification and semantic segmentation. While most efforts in the geospatial domain have focused on developing large models trained from scratch using massive Earth Observation datasets, an alternative strategy that remains underexplored is the reuse and combination of existing pretrained models. In this study, we investigate whether foundation models pretrained on remote sensing and general vision datasets can be effectively combined to improve performance across a diverse set of key Earth Observation tasks. Using the GEO-Bench benchmark, we evaluate several prominent models, including Prithvi, Hiera, and DOFA, on eleven datasets covering a range of spatial resolutions, sensor modalities, and task types. The results show that feature-level ensembling of smaller pretrained models can match or exceed the performance of much larger models, while requiring less training time and computational resources. Moreover, the study highlights the potential of applying knowledge distillation to transfer the strengths of ensembles into more compact models, offering a practical path for deploying foundation models in real-world Earth Observation applications.
Abstract (translated)
基础模型正在通过提供通用且可扩展的解决方案，如场景分类和语义分割等关键任务，迅速改变地球观测数据挖掘领域。尽管地理空间领域的大部分工作主要集中在开发大型模型上，这些模型从大规模地球观测数据集中从头开始训练，但一种尚未充分探索的方法是重用和组合现有的预训练模型。在这项研究中，我们调查了预先在遥感和通用视觉数据集上进行训练的基础模型是否可以有效地结合使用，以改善一系列重要地球观测任务的性能表现。 利用GEO-Bench基准测试，我们在包括Prithvi、Hiera和DOFA在内的几个突出模型上进行了评估，这些模型覆盖了多种空间分辨率、传感器模式及任务类型，共在11个数据集上进行验证。结果显示，较小预训练模型的功能级集成可以与较大模型匹配或超越其性能，并且所需的训练时间和计算资源更少。 此外，该研究还强调了将知识蒸馏应用于将集成的优势转移到更加紧凑的模型中的潜力，为实际部署基础模型在地球观测应用中提供了一条实用路径。
URL
https://arxiv.org/abs/2506.20174
