Related Experiment Video
Updated: Jun 5, 2026
07:13
Early Detection of Cyanobacterial Blooms and Associated Cyanotoxins using Fast Detection Strategy
Published on: February 25, 2021
4.3K
RivAIrSet: A multitemporal high-resolution UAV imagery dataset for machine learning-based river water segmentation
Marco La Salandra1, Rosa Colacicco1, Pierfrancesco Dellino1
1Department of Earth and Geoenvironmental Sciences, University of Bari, Via Orabona 4, 70125 Bari, Italy.
Data in Brief
|January 5, 2026
Summary
Researchers developed RivAIrSet, a dataset of 7630 Unmanned Aerial Vehicle (UAV) images of the Basento River. This resource aids machine learning for river water segmentation and hydrological model development.
Area of Science:
- Earth and Environmental Sciences
- Computer Science
Background:
- River dynamics monitoring requires frequent, systematic observations.
- Machine learning (ML) is increasingly used for spatiotemporal mapping of fluvial environments.
- Robust ML models necessitate large, well-annotated, multitemporal datasets.
Purpose of the Study:
- Introduce RivAIrSet, a novel dataset for river research.
- Facilitate ML-based river water segmentation and hydrological model calibration.
- Promote open data sharing and comparative studies in UAV-based river monitoring.
Main Methods:
- Collected 7630 high-resolution RGB images using Unmanned Aerial Vehicles (UAVs).
- Acquired images along the Basento River (Southern Italy) under diverse conditions.
- Annotated images to delineate river water areas for ML training and validation.
Main Results:
- Established RivAIrSet, a comprehensive dataset of river imagery.
- Captured variations in hydrological and meteorological conditions.
- Provided annotated data crucial for ML applications in fluvial studies.
Conclusions:
- RivAIrSet is a valuable resource for advancing ML in fluvial environments.
- The dataset supports the development of intelligent river monitoring systems.
- The accompanying open repository fosters collaborative research and data accessibility.
