A robust single-pixel particle image velocimetry based on fully convolutional networks with cross-correlation embedded

Gao, Qi; Lin, Hongtao; Tu, Han Yen; Zhu, Haoran; Wei, Runjie; Zhang, Guoping; Shao, Xueming

doi:10.1063/5.0077146

Cited by 25 publications

(11 citation statements)

References 42 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The main challenges associated with CNN-based methods pertain to their generalization and robustness in the presence of noise. To address these issues, Gao et al [47] combined the capabilities of CNNs in achieving high resolution with the robustness of CC against noise. The method, referred to as CC-FCN, embodies two inputs, i.e., the particle images and low-resolution field obtained by CC.…”

Section: Extraction Of Velocity Fields Using Neural Networkmentioning

confidence: 99%

Machine learning for flow field measurements: a perspective

Discetti¹,

Liu²

2022

Meas. Sci. Technol.

View full text Add to dashboard Cite

Advancements in machine-learning (ML) techniques are driving a paradigm shift in image processing. Flow diagnostics with optical techniques is not an exception. Considering the existing and foreseeable disruptive developments in flow field measurement techniques, we elaborate this perspective, particularly focused to the field of particle image velocimetry. The driving forces for the advancements in ML methods for flow field measurements in recent years are reviewed in terms of image preprocessing, data treatment and conditioning. Finally, possible routes for further developments are highlighted.

show abstract

Section: Extraction Of Velocity Fields Using Neural Networkmentioning

confidence: 99%

Machine learning for flow field measurements: a perspective

Discetti¹,

Liu²

2022

Meas. Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…The model applies a more powerful CNN and drastically improves the accuracy of the deep learning estimator for PIV. Gao et al [17] improved the neural network in another way. Their work combines crosscorrelation methods and deep learning methods.…”

Section: Existing Work On Pivmentioning

confidence: 99%

“…Gao et al. [17] improved the neural network in another way. Their work combines cross‐correlation methods and deep learning methods.…”

Section: Existing Work On Pivmentioning

confidence: 99%

Unsupervised learning on particle image velocimetry with embedded cross‐correlation and divergence‐free constraint

Chong

Liang

Chen

et al. 2022

IET Cyber-Syst and Robotics

View full text Add to dashboard Cite

Particle image velocimetry (PIV) is an essential method in experimental fluid dynamics. In recent years, the development of deep learning-based methods has inspired new approaches to tackle the PIV problem, which considerably improves the accuracy of PIV. However, the supervised learning of PIV is driven by large volumes of data with ground truth information. Therefore, the authors consider unsupervised PIV methods. There has been some work on unsupervised PIV, but they are not nearly as effective as supervised learning PIV. The authors try to improve the effectiveness and accuracy of unsupervised PIV by adding classical PIV methods and physical constraints. In this paper, the authors propose an unsupervised PIV method combined with the cross-correlation method and divergence-free constraint, which obtains better performance than other unsupervised PIV methods. The authors compare some classical PIV methods and some deep learning methods, such as LiteFlowNet, LiteFlowNet-en, and UnLiteFlowNet with the authors' model on the synthetic dataset. Besides, the authors contrast the results of LiteFlowNet, UnLiteFlowNet and the authors' model on experimental particle images. As a result, the authors' model shows comparable performance with classical PIV methods as well as supervised PIV methods and outperforms the previous unsupervised PIV method in most flow cases. K E Y W O R D S neural network, particle image velocimetry, unsupervised learning | INTRODUCTIONParticle Image Velocimetry (PIV) is one of the most essential measurement methods in experimental fluid mechanics. PIV is a non-intrusive method to detect the fluid flow, so it is widely used in optical, quantitative, and non-contact surface or fluid motion analysis [1]. To conduct a PIV experiment, some tiny particles are necessary. These particles are assumed to follow the flow dynamics. By the illumination of lasers, the particles in the fluid are visible. We can infer the velocity field by comparing the particles in two images, which are generated in a short time interval [2].There are two main techniques used for performing classical PIV: cross-correlation and variational optical flow methods. Cross-correlation methods could provide a sparse motion field by searching the maximum of the crosscorrelation between two interrogation windows of an image pair [3]. The window deformation iterative multi-gridThis is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

show abstract

“…Güemes et al [35] recently introduced a new concept of Randomly Seeded super-resolution GANs (RaSeedGAN) that achieves similar resolution enhancement and does not need a paired low-high resolution dataset for training as it exploits directly the sparse particle measurements as a target. The main drawback of neural-network-based methods is that they require experienced users for training, and the uncertainty quantification is still difficult, as we can see in the novel methodologies proposed by Wang et al [36,37] and Gao et al [38].…”

Section: Introductionmentioning

confidence: 99%