A 3D‐Based Method for Flood Depth Estimation From Social Media Images
Water Resources Research · 2026-10 · 62 (10)
DOI: 10.1029/2025wr043163 ↗Kurz gefasst
Fachartikel in der Zeitschrift Water Resources Research, online erschienen am 12.10.2026. Verfasst von Miao Shen und 2 weiteren Autor:innen. Die Arbeit nennt 70 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Flood depth estimation from social media images provides timely and cost‐effective observations, offering significant potential for flood monitoring and emergency response. Existing image‐based methods typically rely on humans, vehicles, or traffic signs as scale references, however, such objects are frequently absent in real flood scenes and often yield unreliable scale information. In contrast, urban 3D models offer consistent and accurate metric references across large spatial extents. In this paper, we propose a flood depth estimation method based on urban 3D models and develop an Object‐level Coarse‐to‐Fine Matching Model that transfers metric scale from 3D models to social media images. The results demonstrate that our model significantly outperforms the representative 2D3D‐MATR on an urban public benchmark, increasing the inlier ratio by approximately 30% and improving the registration recall by about 16%. Building upon the matching capability, our 3D‐based flood depth estimation method generalizes across diverse flood scenarios and achieves improved accuracy, with a Mean Absolute Error (MAE) of 0.164 m and a Root Mean Square Error (RMSE) of 0.212 m. By transforming discrete social media data into spatially explicit, high‐accuracy flood depth maps, our method not only provides reliable support for rapid flood assessment and emergency decision‐making during events but also offers valuable data for calibrating hydraulic models post‐disaster.
Zitieren
APA
Shen, M., Zhang, F., & Shi, Y. (2026). A 3D‐Based Method for Flood Depth Estimation From Social Media Images. Water Resources Research, 62(10). https://doi.org/10.1029/2025wr043163
Vancouver
Shen M, Zhang F, Shi Y. A 3D‐Based Method for Flood Depth Estimation From Social Media Images. Water Resources Research. 2026;62(10). doi:10.1029/2025wr043163
BibTeX
@article{Shen2026,
author = {Shen, Miao and Zhang, Feng and Shi, Yongguo},
title = {A 3D‐Based Method for Flood Depth Estimation From Social Media Images},
journal = {Water Resources Research},
year = {2026},
volume = {62},
number = {10},
doi = {10.1029/2025wr043163}
}
EUVIMED