Showing 1–1 of 1 results for author: Sohn, D K
-
Three-dimensional particle tracking velocimetry using shallow neural network for real-time analysis
Authors:
Yeonghyeon Gim,
Dong Kyu Jang,
Dong Kee Sohn,
Hyoungsoo Kim,
Han Seo Ko
Abstract:
Three-dimensional particle tracking velocimetry (3D-PTV) technique is widely used to acquire the complicated trajectories of particles and flow fields. It is known that the accuracy of 3D-PTV depends on the mapping function to reconstruct three-dimensional particles locations. The mapping function becomes more complicated if the number of cameras is increased and there is a liquid-vapor interface,…
▽ More
Three-dimensional particle tracking velocimetry (3D-PTV) technique is widely used to acquire the complicated trajectories of particles and flow fields. It is known that the accuracy of 3D-PTV depends on the mapping function to reconstruct three-dimensional particles locations. The mapping function becomes more complicated if the number of cameras is increased and there is a liquid-vapor interface, which crucially affect the total computation time. In this paper, using a shallow neural network model (SNN), we dramatically decrease the computation time with a high accuracy to successfully reconstruct the three-dimensional particle positions, which can be used for real-time particle detection for 3D-PTV. The developed technique is verified by numerical simulations and applied to measure a complex solutal Marangoni flow patterns inside a binary mixture droplet.
△ Less
Submitted 2 February, 2020;
originally announced February 2020.
