Simantiris et al. (2026) Unsupervised Estimation of Post-Event Standing Urban Floodwater Depth Using Aerial Imagery and Digital Terrain Models

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-08-09
• Authors: Georgios Simantiris, Konstantinos Bacharidis, Costas Panagiotakis
• DOI: 10.3390/rs18162673

Research Groups

Not specified

Short Summary

The study proposes an unsupervised, training-free framework that estimates residual floodwater depth by combining color-based segmentation of UAV imagery with Digital Terrain Models (DTMs) based on the hydrostatic equilibrium principle.

Objective

• To develop a rapid, scalable, and unsupervised method for estimating floodwater depth that bypasses the need for labeled training data or computationally intensive hydrodynamic models.

Study Configuration

• Spatial Scale: 12 urban and peri-urban sites in the Southeastern United States.
• Temporal Scale: Post-event analysis of Hurricanes Matthew and Florence.

Methodology and Data

• Models used: Unsupervised color-based segmentation algorithm; Hydrostatic equilibrium principle.
• Data sources: UAV imagery, Digital Terrain Models (DTMs), and the Inundation2Depth dataset.

Main Results

• Segmentation Accuracy: F1-scores for flood extent delineation ranged from 63% to 96%.
• Depth Accuracy: Absolute flood depth Root Mean Square Error (RMSE) ranged from 0.16 m (well-defined catchments) to 1.69 m (highly obscured regions).
• Efficiency: The framework executes within seconds on standard CPU hardware.

Contributions

• Introduces a training-free alternative to supervised deep learning and hydrodynamic models, enabling rapid, first-order flood depth mapping for time-critical emergency response without requiring manual annotations.

Funding

Not specified

Citation

@article{Simantiris2026Unsupervised,
  author = {Simantiris, Georgios and Bacharidis, Konstantinos and Panagiotakis, Costas},
  title = {Unsupervised Estimation of Post-Event Standing Urban Floodwater Depth Using Aerial Imagery and Digital Terrain Models},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18162673},
  url = {https://doi.org/10.3390/rs18162673}
}