Three-dimensional measurement of the droplets out of focus in shadowgraphy systems via deep learning-based image-processing method

Article Computer Science, Artificial Intelligence

Defocus Image Deblurring Network With Defocus Map Estimation as Auxiliary Task

Haoyu Ma et al.

Summary: This paper proposes a network architecture called DID-ANet for single image defocus deblurring by using defocus map estimation as an auxiliary task. A large-scale dataset is also built for network training.

IEEE TRANSACTIONS ON IMAGE PROCESSING (2022)

添加到收藏夹

Article Mechanics

How surface roughness promotes or suppresses drop splash

Haixiang Zhang et al.

Summary: This work presents an experimental study on the splashing behaviors of drops impinging on rough solid surfaces. It is found that the increase in surface roughness can promote splash of low-viscosity drops but suppress splash of high-viscosity drops. The splash regime can be classified into corona splash and prompt splash, and the increase in surface roughness can promote the prompt splash but suppress the corona splash.

PHYSICS OF FLUIDS (2022)

添加到收藏夹

Article Mechanics

Penetration and aerosolization of cough droplet spray through face masks: A unique pathway of transmission of infection

Gautham Vadlamudi et al.

Summary: This study investigates the effectiveness of different types of masks in blocking droplet transmission. It shows that masks are not able to completely prevent the penetration of larger droplets and can cause atomization, leading to increased risk of virus transmission. The study also examines the filtration and entrapment of virus-like particles in masks, as well as their efficacy for susceptible individuals.

PHYSICS OF FLUIDS (2022)

添加到收藏夹

Article Mechanics

Deep-learning-based super-resolution reconstruction of high-speed imaging in fluids

Zhibo Wang et al.

Summary: A degradation and super-resolution attention model using unsupervised machine learning is proposed to reconstruct high resolution time-resolved fluid images from coarse data, validated in high-speed experiments. The model outperforms existing methods and has significant applications in obtaining flow images with high spatial and temporal resolution.

PHYSICS OF FLUIDS (2022)

添加到收藏夹

Article Mechanics

BubDepth: A neural network approach to three-dimensional reconstruction of bubble geometry from single-view images

Chaoyue Gong et al.

Summary: This paper proposes a novel method for the 3D reconstruction of bubble shape based on single-view images. By utilizing grayscale information and neural networks, it can automatically reconstruct the rough 3D shapes of one side of bubbles in the images.

INTERNATIONAL JOURNAL OF MULTIPHASE FLOW (2022)

添加到收藏夹

Article Engineering, Multidisciplinary

Machine learning shadowgraph for particle size and shape characterization

Jiaqi Li et al.

Summary: This paper introduces a robust learning-based method using a single convolution neural network for analyzing particle shadow images. The method employs a two-channel-output U-net model to generate binary particle image and particle centroid image, and achieves better performance compared to traditional methods through marker-controlled watershed segmentation. The proposed machine learning approach shows great potential for real-time particle image analysis in industrial applications.

MEASUREMENT SCIENCE AND TECHNOLOGY (2021)

添加到收藏夹

Article Engineering, Mechanical

Sensitivity analysis and measurement uncertainties of a two-camera depth from defocus imaging system

Wu Zhou et al.

Summary: This study comprehensively investigated the impact of system and particle parameters on the calibration functions and measurement uncertainties of a two-camera Depth from Defocus imaging system, with experimental verification. It provides explicit recommendations on which quantities are most influential in determining measurement accuracy and must be carefully controlled.

EXPERIMENTS IN FLUIDS (2021)

添加到收藏夹

Article Engineering, Mechanical

Motion blur treatment utilizing deep learning for time-resolved particle image velocimetry

Jeong Suk Oh et al.

Summary: The study introduces a new method for reducing motion blur-induced errors in time-resolved particle image velocimetry. By applying the Monte-Carlo method to synthetic images, it was found that large streaks cause significant errors and outliers in displacement estimation. A novel deblur filter using a generative adversarial network was developed, effectively reducing errors and outliers in testing scenarios.

EXPERIMENTS IN FLUIDS (2021)

添加到收藏夹

Article Mechanics

Reversed role of liquid viscosity on drop splash

Haixiang Zhang et al.

Summary: The influence of liquid viscosity on the splashing behavior of a drop impacting a smooth surface was investigated. It was found that liquid viscosity plays a reversed role on the drop splash, promoting splash in low-viscosity cases and suppressing splash in high-viscosity cases. The increase in liquid viscosity contributes to a lower splashing angle and influences both aerodynamic and viscous forces, resulting in a proposed splash criterion for smooth surfaces.

PHYSICS OF FLUIDS (2021)

添加到收藏夹

Article Mechanics

Experimental velocity data estimation for imperfect particle images using machine learning

Masaki Morimoto et al.

Summary: This study introduces a method using supervised machine learning to estimate velocity fields from particle images with missing regions due to experimental limitations. The machine learning models show promising results in estimating velocity fields in missing regions compared to conventional methods, with significantly higher data density.

PHYSICS OF FLUIDS (2021)

添加到收藏夹

Article Engineering, Multidisciplinary