2026-07-21T14:18:16Z https://drops.dagstuhl.de/oai

oai:drops-oai.dagstuhl.de:23847 2025-11-12T13:20:30Z ddc:004 open_access

U-Prithvi: Integrating a Foundation Model and U-Net for Enhanced Flood Inundation Mapping Kostejn, Vit Essus, Yamil Abrahamson, Jenna Vatsavai, Ranga Raju GeoAI flood mapping foundation model U-Net Prithvi In recent years, large pre-trained models, commonly referred to as foundation models, have become increasingly popular for various tasks leveraging transfer learning. This trend has expanded to remote sensing, where transformer-based foundation models such as Prithvi, msGFM, and SatSwinMAE have been utilized for a range of applications. While these transformer-based models, particularly the Prithvi model, exhibit strong generalization capabilities, they have limitations on capturing fine-grained details compared to convolutional neural network architectures like U-Net in segmentation tasks. In this paper, we propose a novel architecture, U-Prithvi, which combines the strengths of the Prithvi transformer with those of U-Net. We introduce a RandomHalfMaskLayer to ensure balanced learning from both models during training. Our approach is evaluated on the Sen1Floods11 dataset for flood inundation mapping, and experimental results demonstrate better performance of U-Prithvi over both individual models, achieving improved performance on out-of-sample data. While this principle is illustrated using the Prithvi model, it is easily adaptable to other foundation models. Schloss Dagstuhl – Leibniz-Zentrum für Informatik Vit Kostejn and Yamil Essus and Jenna Abrahamson and Ranga Raju Vatsavai 2025 Is Part Of LIPIcs, Volume 346, 13th International Conference on Geographic Information Science (GIScience 2025) InProceedings Text doc-type:ResearchArticle publishedVersion application/pdf doi:10.4230/LIPIcs.GIScience.2025.18 urn:nbn:de:0030-drops-238479 https://drops.dagstuhl.de/entities/document/10.4230/LIPIcs.GIScience.2025.18 eng https://creativecommons.org/licenses/by/4.0/legalcode