Search research articles

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

Related Concept Videos

Typical Model Studies

Typical Model Studies

Fluid mechanics model studies often utilize scaled-down systems to predict fluid behavior in full-scale environments, such as river flows, dam spillways, and structures interacting with open surfaces. Maintaining Froude number similarity in river models is crucial, as it replicates surface flow features like wave patterns and velocities.

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Evaluating Molecular Representations for Predicting Cyclodextrin-PFAS Binding Energy with Machine Learning: Domain Transfer and Data Limitations.

Journal of chemical information and modeling·2026

Same author

Exploring transferability of plastic-water hyacinth interaction and detection in rivers.

iScience·2026

Same author

Three rivers to the test: Lessons for global plastic transport modelling.

Environmental pollution (Barking, Essex : 1987)·2026

Same author

A topology-derived flow inference approach to optimize sensor placement for effective inflow and infiltration detection in sewer networks.

Water research·2026

Same author

Exploring plastic detectability on riverbanks using remote sensing.

Marine pollution bulletin·2026

Same author

A global dataset of microplastics in urban stormwater runoff (2018-2024).

Scientific data·2026

Same journal

Non-point source pollution prediction and dynamics simulation in urban runoff: a physics-informed neural network approach.

Water research·2026

Same journal

A multivariate water quality forecasting model with dynamic variable selection and dissolved oxygen physical-consistency constraints.

Water research·2026

Same journal

Spectroscopic and multivariate analysis insights into the DOM composition and distribution in a full-scale photovoltaic wastewater zero-liquid discharge system.

Water research·2026

Same journal

Assessing climate change-induced precipitation effects on the long-term reliability of common final landfill cover systems.

Water research·2026

Same journal

Synergetic piezo-photocatalytic tuning internal electric field in S-scheme heterojunction for highly efficient tetracycline degradation and hydrogen production.

Water research·2026

Same journal

Mimosa-inspired dual-responsive flocculant with controllable hydrophobicity/hydrophilicity reconstruction for the purification of triclosan containing emulsified wastewater.

Water research·2026

See all related articles

Search research articles

Related Experiment Video

Updated: Jan 13, 2026

Watershed Planning within a Quantitative Scenario Analysis Framework

Watershed Planning within a Quantitative Scenario Analysis Framework

Published on: July 24, 2016

A semi-supervised learning-based framework for quantifying litter fluxes in river systems.

Tianlong Jia¹, Riccardo Taormina², Rinze de Vries³

¹Delft University of Technology, Faculty of Civil Engineering and Geosciences, Department of Water Management, Stevinweg 1, 2628 CN Delft, The Netherlands; Karlsruhe Institute of Technology (KIT), Institute of Water and Environment, Karlsruhe, Germany.

|October 29, 2025

Summary

This summary is machine-generated.

A new semi-supervised learning (SSL) framework improves floating plastic detection in rivers, outperforming traditional methods. This approach enhances river pollution monitoring by better identifying small litter items and quantifying fluxes.

Keywords:

Artificial intelligence Camera images Contrastive learning Environmental monitoring Macroplastic flux Object detection Pollution SwAV

More Related Videos

Automatic Image Processing to Determine the Community Size Structure of Riverine Macroinvertebrates

Automatic Image Processing to Determine the Community Size Structure of Riverine Macroinvertebrates

Published on: January 13, 2023

Image-based Lagrangian Particle Tracking in Bed-load Experiments

Image-based Lagrangian Particle Tracking in Bed-load Experiments

Published on: July 20, 2017

Related Experiment Videos

Last Updated: Jan 13, 2026

Watershed Planning within a Quantitative Scenario Analysis Framework

Watershed Planning within a Quantitative Scenario Analysis Framework

Published on: July 24, 2016

Automatic Image Processing to Determine the Community Size Structure of Riverine Macroinvertebrates

Automatic Image Processing to Determine the Community Size Structure of Riverine Macroinvertebrates

Published on: January 13, 2023

Image-based Lagrangian Particle Tracking in Bed-load Experiments

Image-based Lagrangian Particle Tracking in Bed-load Experiments

Published on: July 20, 2017

Area of Science:

Environmental Science
Computer Science
Remote Sensing

Background:

Supervised deep learning (SL) is used for detecting floating macroplastic litter in water bodies.
Quantifying litter fluxes in wide rivers is crucial for pollution assessment but understudied.
SL models require extensive labeled data and struggle with small litter detection.

Purpose of the Study:

To propose a semi-supervised learning (SSL)-based framework combined with Slicing Aided Hyper Inference (SAHI) for quantifying cross-sectional floating litter fluxes in rivers.
To overcome limitations of SL methods, including data requirements and detection of small litter items.

Main Methods:

A four-step framework: image collection, SSL model development, SAHI application for detection, and flux quantification.
SSL involves self-supervised pre-training on unlabeled data and supervised fine-tuning on limited labeled data using Faster R-CNN with ResNet50 backbone.
Evaluation of in-domain and zero-shot out-of-domain detection performance, and flux quantification accuracy compared to SL and human counting.

Main Results:

SSL models showed improved in-domain (F1-score +0.2) and out-of-domain (+0.14) detection compared to SL benchmarks, benefiting from larger pre-training datasets and epochs.
SAHI improved F1-score by up to 0.19 by identifying 45 additional small litter items (<1,000 cm²).
The SSL framework underestimated fluxes by 3-4x compared to human measurements but estimated twice the fluxes of the baseline SL framework.

Conclusions:

The proposed SSL-based framework with SAHI shows significant potential for enhancing floating litter detection and flux measurement in rivers.
SSL models offer improved performance over SL methods, especially with sufficient pre-training data.
Further scaling with broader datasets can advance global litter monitoring systems, despite current underestimation of fluxes due to missed transparent or entangled items.