High Resolution Flood Extent Detection Using Deep Learning with Random Forest Derived Training Labels

Azizbek Nuriddinov; Ebrahim Ahmadisharaf; Mohammad Reza Alizadeh

cs.CVMar 23, 2026

High Resolution Flood Extent Detection Using Deep Learning with Random Forest Derived Training Labels

Authors: Azizbek Nuriddinov, Ebrahim Ahmadisharaf, Mohammad Reza Alizadeh

Organizations: Department of Civil Engineering, Florida State University, Tallahassee, FL, USA · Dept. of Biosystems & Agricultural Engineering, Michigan State University, East Lansing, MI, USA

Abstract

Validation of flood models, used to support risk mitigation strategies, remains challenging due to limited observations during extreme events. High-frequency, high-resolution optical imagery (~3 m), such as PlanetScope, offers new opportunities for flood mapping, although applications remain limited by cloud cover and the lack of labeled training data during disasters. To address this, we develop a flood mapping framework that integrates PlanetScope optical imagery with topographic features using machine learning (ML) and deep learning (DL) algorithms. A Random Forest model was applied to expert-annotated flood masks to generate training labels for DL models, U-Net. Two U-Net models with ResNet18 backbone were trained using optical imagery only (4 bands) and optical imagery combined with Height Above Nearest Drainage (HAND) and topographic slope (6 bands). Hurricane Ida (September 2021), which caused catastrophic flooding across the eastern United States, including the New York City metropolitan area, was used as an example to evaluate the framework. Results demonstrate that the U-Net model with topographic features achieved very close performance to the optical-only configuration (F1=0.92 and IoU=0.85 by both modeling scenarios), indicating that HAND and slope provide only marginal value to inundation extent detection. The proposed framework offers a scalable and label-efficient approach for mapping inundation extent that enables modeling under data-scarce flood scenarios.

Figures & tables

Fig. 1: Study area map showing the HUC-6 Lower Hudson Basin in the northeastern United States.

Fig. 2: Methodological framework flowchart depicting the three-stage workflow. Stage 1 - Data Aggregation: PlanetScope Composite Tiles (RGB + NIR) combined with HAND, Slope, and HWM reference points. Stage 2 - Preparing Annotations with Random Forest: A 1024 $\times$ 1024 $\times$ 7 band cube (RGB + NIR + Slope + HAND + NDWI) containing handcrafted expert annotation serves as RF training input, producing RF predictions and flood masks tested on all samples. Stage 3 - Deep Learning for Flood Prediction: Multiple 128 $\times$ 128 $\times$ 7 training crops are used as annotations for training the deep learning models. UNet with ResNet18 backbone performs training and testing, generating final flood mask predictions on all composite tiles.

Model	Bands	Train/Test	IoU	F1
Random Forest	RGB, NIR, Slope,	50% pixels,	0.708	0.829
	HAND, NDWI	handcrafted tile
UNet (4-band)	RGB, NIR	768/320 crops	0.8507	0.9193
		(128 $\times$ 128)
UNet (6-band)	RGB, NIR,	768/320 crops	0.8520	0.9201
	Slope, HAND	(128 $\times$ 128)

TABLE I: Performance Comparison of Different Learning Algorithms

Fig. 3: Flood detection method comparison on PlanetScope imagery. (A) Original satellite image, (B) NDWI thresholded results showing false positives (red circles) and false negatives (yellow squares), (C) Random Forest predictions with fewer false positives but patchy results, (D) Deep Learning 4-band with reduced noisy detections, and (E) Deep Learning 6-band achieving best results with reduced noise. The progression demonstrates improvement from traditional index-based methods to machine learning and deep learning approaches.

Flood Inundation Mapping

Flood inundation mapping uses satellite imagery, terrain data, and models to locate flooded areas and estimate water depth for response and planning.

See all 27 papers

Flood Mapping from RGB imagery using a Vision Foundation Model
Jun 23, 2026Vladyslav Polushko, Tilman Bucher, Ronald Rösch +3Remote Sensing Image Segmentation Geospatial Foundation Models
Image Processing Department, Fraunhofer ITWM, Kaiserslautern, Germany · Institut für Optische Sensorsysteme, DLR, Berlin, Germany · ACIDA Lab, Hochschule Darmstadt, Darmstadt, Germany +1
High-Resolution Flood Mapping With Sentinel-1 and Sentinel-2 via Misalignment-Robust Cross-Sensor Learning and Generative Despeckling
Jun 29, 2026David Ma, Jeremy Feinstein, Shreya Pandit +2Remote Sensing Image Segmentation Flood Inundation Mapping
Argonne National Laboratory, Lemont, IL 60439 USA · Ohlone College, Fremont, CA 94539 USA
Deep Learning for Remote Sensing to Improve Flood Inundation Mapping
Jun 1, 2026Yogesh Bhattarai, Vijay Chaudhary, Wai Lim Kim +1Diffusion Models for Image Restoration Flood Inundation Mapping
Department of Civil and Environmental Engineering Howard University Washington, D.C, 20060, USA. · Department of Computer Science and Electrical Engineering Howard University Washington, D.C, 20060, USA. · Department of Medicine Howard University Washington, D.C, 20060, USA.

Jun 23, 2026J/K move · Enter open · S save

Flood Mapping from RGB imagery using a Vision Foundation Model

Vladyslav Polushko, Tilman Bucher, Ronald Rösch, Thomas März+2

Image Processing Department, Fraunhofer ITWM, Kaiserslautern, Germany · Institut für Optische Sensorsysteme, DLR, Berlin, Germany · ACIDA Lab, Hochschule Darmstadt, Darmstadt, Germany+1

Remote Sensing Image Segmentation Geospatial Foundation Models Domain Generalization Vision Foundation Model Adaptation Flood Inundation Mapping Remote Sensing

Open paper AskPDF arXiv

Timely, high-resolution maps of flood extent around settlements are essential for emergency response and damage assessment. We consider airborne RGB imagery for flood mapping as it can be collected rapidly at low cost. To produce flood maps, deep learning models for water segmentation are often used. CNN based and small vision transformer models are used. However, they need much data for adaptation to a change of scenery, i.e., another flooding event. Vision foundation models or large vision transformers are known to generalize across domains. Recently, foundation models for Earth observation became available. They are pretrained on satellite data, whose spatial resolution, viewing geometry, and radiometry differ from nadir RGB imagery. Thus, adaptation is required. We investigate how a satellite-pretrained Earth observation foundation model can be adapted to centimeter-scale floodwater mapping from RGB imagery. Specifically, we fine-tune a model we call Prithvi-2.0-UPN consisting of the Prithvi-EO-2.0-600M Vision Transformer combined with a UPerNet decoder for binary water segmentation on two RGB datasets (BlessemFlood21, NeuenahrFlood). In a first experiment we observe that Prithvi-2.0-UPN reaches state-of-the-art results on BlessemFlood21 and NeuenahrFlood, when trained on their datasets. In a second experiment we show that Prithvi-2.0-UPN performs better than state-of-the-art baseline models for transfer to a new flood event (trained on BlessemFlood21, tested on NeuenahrFlood) in a zero-shot setting. However, the performance indicates room for improvement. In this respect, we investigate in a third experiment how performance improves when further fine-tuning the models with small shares of NeuenahrFlood training data: Prithvi-2.0-UPN improves the fastest and reaches almost the performance level when fully trained on NeuenahrFlood, indicating transfer capabilities.

Abstract

Flood Inundation Mapping

Explore similar work

Flood Mapping from RGB imagery using a Vision Foundation Model

High-Resolution Flood Mapping With Sentinel-1 and Sentinel-2 via Misalignment-Robust Cross-Sensor Learning and Generative Despeckling

Deep Learning for Remote Sensing to Improve Flood Inundation Mapping