Assessing Inundation Semantic Segmentation Models Trained on High- versus Low-Resolution Labels using FloodPlanet, a Manually Labeled Multi-Sourced High-Resolution Flood Dataset
Journal of Remote Sensing · 2025-01 · 5
DOI: 10.34133/remotesensing.0575 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Remote Sensing, online erschienen am 15.05.2025. Verfasst von Zhijie Zhang und 9 weiteren Autor:innen. Die Arbeit wurde bisher 16-mal zitiert (Crossref) und nennt 61 Literaturangaben. Gefördert von: NASA Commercial Smallsat Data Acquisition Program.
Abstract (Originalsprache)
Flooding impacts more people than any other environmental hazard, causing extensive economic and social impact. Leveraging satellite data and deep learning substantially improves flood monitoring and, potentially, management. However, deep learning efforts are frequently constrained by the limited availability of high-quality training and validation datasets, as well as the resolution, spatial coverage, and temporal limitations of inundation observation from public sensors. To address these challenges and contribute to the future development of geo-foundation models requiring extensive data and validation, we curate and publicly release FloodPlanet, a manually labeled inundation dataset [ 1 ]. FloodPlanet stands out for its manual annotations based on 3-m high-resolution commercial data from PlanetScope, and diverse ecoregions and flood event coverage. The dataset includes 366 hand-annotated labels, each 1,024 × 1,024 pixels, with corresponding Sentinel-1 and Sentinel-2 imagery, covering 19 global flood events from 2017 to 2020. Employing a “leave-one-region-out” cross-validation approach with a baseline UNet model, we achieved a mean intersection over union score (IoU) of 0.691 (SD: 0.227) for inundation detection across all events with PlanetScope, which is around 20% higher compared to Sentinel-1 and Sentinel-2 from the same event. Comparative analysis using PlanetScope labels to train models with Sentinel-1 and Sentinel-2 data revealed that FloodPlanet labels improve public sensor-based inundation detection by up to 15.6% (SD: 0.242) in IoU. These results imply that even if commercial data are too costly for near real-time inference applications, using some commercial data to train public sensor models could be an important lower-cost investment to increase accuracy.
Förderung
- NASA Commercial Smallsat Data Acquisition Program
Zitieren
APA
Zhang, Z., Giezendanner, J., Mukherjee, R., Tellman, B., Melancon, A., Purri, M., Gurung, I., Lall, U., Barnard, K., & Molthan, A. (2025). Assessing Inundation Semantic Segmentation Models Trained on High- versus Low-Resolution Labels using FloodPlanet, a Manually Labeled Multi-Sourced High-Resolution Flood Dataset. Journal of Remote Sensing, 5. https://doi.org/10.34133/remotesensing.0575
Vancouver
Zhang Z, Giezendanner J, Mukherjee R, Tellman B, Melancon A, Purri M, et al. Assessing Inundation Semantic Segmentation Models Trained on High- versus Low-Resolution Labels using FloodPlanet, a Manually Labeled Multi-Sourced High-Resolution Flood Dataset. Journal of Remote Sensing. 2025;5. doi:10.34133/remotesensing.0575
BibTeX
@article{Zhang2025,
author = {Zhang, Zhijie and Giezendanner, Jonathan and Mukherjee, Rohit and Tellman, Beth and Melancon, Alexander and Purri, Matt and Gurung, Iksha and Lall, Upmanu and Barnard, Kobus and Molthan, Andrew},
title = {Assessing Inundation Semantic Segmentation Models Trained on High- versus Low-Resolution Labels using FloodPlanet, a Manually Labeled Multi-Sourced High-Resolution Flood Dataset},
journal = {Journal of Remote Sensing},
year = {2025},
volume = {5},
doi = {10.34133/remotesensing.0575}
}
EUVIMED