Work

Bubble Evolution Detector B.E.D. – A Neural Network-Based Approach to Accurately Detect, Classify, and Evaluate Gas Bubbles Captured by A High-Speed Camera on Textured Surfaces

Year: 2025

Type: article

Abstract: Gas bubble emergence is an important indicator of the performance in many processes, for example electrochemical reactions. Using a convolutional neural network (CNN) based on the Darknet/YOLO4-architecture; this software allows the detection and tracking of gas bubbles from high-speed camera videos, even on strongly textured backgrounds. Further, it evaluates growth rates, detachment size and merging of gas bubbles. This allows a good assessment of many important gas formation characteristics, thus helping performance evaluation and identify potential for improvement.

Source: Journal of Open Research Software

DOI: 10.5334/jors.505

Authors: Lukas Lentz, Dorian Hüne, Sebastian Handrich, Christoph Niems, Thomas Gimpel

Institutions: Forschungszentrum Energiespeichertechnologien, Clausthal University of Technology

Language: English

Topic: Minerals Flotation and Separation Techniques

Subfield: Water Science and Technology

Field: Environmental Science

Domain: Physical Sciences

Open Access status: gold

Estimated APC paid: $490

Funder: Deutsche Forschungsgemeinschaft