A 3D reconstruction method of bubble flow field based on multi-view images by bi-direction filtering maximum likelihood expectation maximization algorithm

Wang, Hongyi; Yang, Yang; Dou, Gongcheng; Lou, Jipei; Zhu, Xinjun; Song, Limei; Dong, Feng

doi:10.1016/j.ijmultiphaseflow.2023.104480

Cited by 6 publications

(2 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…CNNs have proven to be superior to traditional computer vision methods for image segmentation, achieving high resolution (He et al, 2020;Hessenkemper et al, 2022). Various CNN approaches have been explored for bubble segmentation, such as the use of region-based CNNs like Faster-RCNN (Haas et al, 2020), sliding window-based CNN (Poletaev et al, 2020), and the implementation of segmentation masks through Mask RCNN, which not only classifies but also assigns each pixel to individual bubble objects (Cui et al, 2021;Hessenkemper et al, 2022;Kim & Park, 2021;Wang et al, 2023). Although significant progress made with CNN-based techniques has been made in recent years, challenges persist in segmenting very complex images, such as those obtained in industrial settings, which are typically subject to wide bubble-size distributions, high gas content, cloudiness due to the presence of solid particles, and lighting issues.…”

Section: Introductionmentioning

confidence: 99%

Artificial neural network (ANN) modelling to estimate bubble size from macroscopic image and object features

Vinnett,

León,

Mesa

2024

Physicochem. Probl. Miner. Process.

View full text Add to dashboard Cite

Bubble size measurements in aerated systems such as froth flotation cells are critical for controlling gas dispersion. Commonly, bubbles are measured by obtaining representative photographs, which are then analyzed using segmentation and identification software tools. Recent developments have focused on enhancing these segmentation tools. However, the main challenges around complex bubble cluster segmentation remain unresolved, while the tools to tackle these challenges have become increasingly complex and computationally expensive. In this work, we propose an alternative solution, circumventing the need for image segmentation and bubble identification. An Artificial Neural Network (ANN) was trained to estimate the Sauter mean bubble size (D<sub>32</sub>) based on macroscopic image features obtained with simple and inexpensive image analysis. The results showed excellent prediction accuracy, with a correlation coefficient, R, over 0.998 in the testing stage, and without bias in its error distribution. This machine learning tool paves the way for robust and fast estimation of bubble size under complex bubble images, without the need of image segmentation.

show abstract

Section: Introductionmentioning

confidence: 99%

Artificial neural network (ANN) modelling to estimate bubble size from macroscopic image and object features

Vinnett,

León,

Mesa

2024

Physicochem. Probl. Miner. Process.

View full text Add to dashboard Cite

show abstract

“…In the case of particle size, this problem has been addressed by employing 3D measurement techniques such as 3D X-ray micro-CT [14,15]. Although there exist some 3D techniques available for bubble size measurement [16][17][18][19], these have not been widely adopted in froth flotation or mineral processing due to the multiphase nature of flotation and the practical constraints of using multiple cameras. Most practitioners and researchers continue to measure bubble size using a 2D single-camera approach, and therefore it is critical to quantify the impact of bubble size uncertainties on flotation and reagent assessments.…”

Section: Introductionmentioning

confidence: 99%

The Role of Stereological Assumptions in Bubble Size Estimations and Their Implications for Assessing Critical Coalescence Concentrations

Vinnett

Mesa

2023

Minerals

View full text Add to dashboard Cite

Accurate measurement of bubble size is critical for assessing flotation performance. However, the 3D nature of bubbles, in contrast to the 2D nature of photographs obtained using a bubble viewer apparatus, may lead to distortions related to stereological assumptions. This study aimed to quantify the impact of these stereological effects on bubble size measurements in frother characterisations. Our results showed that different assumptions regarding bubble shape and volume resulted in variations in bubble size calculations of up to 10%. Furthermore, these stereological effects were propagated to the calculation of the critical coalescence concentration, leading to uncertainties of up to 14% depending on the type of frother. These findings emphasise the importance of considering stereological effects and selecting an appropriate calculation method when measuring bubble size for flotation and reagent assessments.

show abstract

Advances in the modeling of multiphase flows and their application in nuclear engineering—A review

Wu,

Zhang,

Gui

et al. 2024

Exp. Comput. Multiph. Flow

View full text Add to dashboard Cite

A 3D reconstruction method of bubble flow field based on multi-view images by bi-direction filtering maximum likelihood expectation maximization algorithm

Cited by 6 publications

References 38 publications

Artificial neural network (ANN) modelling to estimate bubble size from macroscopic image and object features

Artificial neural network (ANN) modelling to estimate bubble size from macroscopic image and object features

The Role of Stereological Assumptions in Bubble Size Estimations and Their Implications for Assessing Critical Coalescence Concentrations

Advances in the modeling of multiphase flows and their application in nuclear engineering—A review

Contact Info

Product

Resources

About