Machine Learning for 3D Particle Tracking in Granular Gases
- PDF / 2,746,718 Bytes
- 10 Pages / 595.224 x 790.955 pts Page_size
- 8 Downloads / 239 Views
ORIGINAL ARTICLE
Machine Learning for 3D Particle Tracking in Granular Gases Dmitry Puzyrev1
· Kirsten Harth1 · Torsten Trittel1 · Ralf Stannarius1
Received: 25 March 2020 / Accepted: 3 May 2020 © The Author(s) 2020
Abstract Dilute ensembles of granular matter (so-called granular gases) are nonlinear systems which exhibit fascinating dynamical behavior far from equilibrium, including non-Gaussian distributions of velocities and rotational velocities, clustering, and violation of energy equipartition. In order to understand their dynamic properties, microgravity experiments were performed in suborbital flights and drop tower experiments. Up to now, the experimental images were evaluated mostly manually. Here, we introduce an approach for automatic 3D tracking of positions and orientations of rod-like particles in a dilute ensemble, based on two-view video data analysis. A two-dimensional (2D) localization of particles is performed using a Mask R-CNN neural network trained on a custom data set. The problem of 3D matching of the particles is solved by minimization of the total reprojection error, and finally, particle trajectories are tracked so that ensemble statistics are extracted. Depending on the required accuracy, the software can work fully self-sustainingly or serve as a base for subsequent manual corrections. The approach can be extended to other 3D and 2D particle tracking problems. Keywords Machine learning · Granular gas · Particle tracking · Object detection · Mask-CNN
Introduction Granular gases are dilute ensembles of macroscopic grains, which in the simplest case interact only upon contact during collisions, without any long-range interactions. Studies of granular gases are relevant, for example, for gaining deeper insights into fundamental physics of non-equilibrium systems (P¨oschel and Luding 2001; P¨oschel and Brilliantov 2003, 2004), as a basis for modeling collisional dynamics in planetary rings or other astrophysical assemblies of solid objects, even in some stages of planet formation (Hestroffer et al. 2019). A quantitative macroscopic description of their ensemble properties will also aid simulations of fluidized granular materials.
Electronic supplementary material The online version of this article (https://doi.org/10.1007/s12217-020-09800-4) contains supplementary material, which is available to authorized users. This article belongs to the Topical Collection: The Effect of Gravity on Physical and Biological Phenomena Guest Editor: Valentina Shevtsova Dmitry Puzyrev
[email protected] 1
Institute of Physics, Otto von Guericke University, Universit¨atsplatz 2, D-39106 Magdeburg, Germany
In contrast to molecular gases, all collisions among particles or between particles and the container walls are dissipative, i.e. part of the kinetic energy is lost. Consequently, without external energy supply, the ensemble will gradually lose its kinetic energy in a process called granular cooling. This cooling process itself is nontrivial, and most of its properties are insufficiently proven
Data Loading...
