Particle reconstruction of volumetric particle image velocimetry with the strategy of machine learning

• Published in: Advances in aerodynamics Vol. 3; no. 1; pp. 1 - 14
• Main Authors: Gao, Qi, Pan, Shaowu, Wang, Hongping, Wei, Runjie, Wang, Jinjun
• Format: Journal Article
• Language: English
• Published: Singapore Springer Nature Singapore 23.09.2021; Springer Nature B.V; SpringerOpen
• Subjects: Aerospace Technology and Astronautics; Algorithms; Convolutional neural network; Engineering; Ghosts; Inverse problems; Machine learning; Neural networks; Particle reconstruction; Volumetric particle image velocimetry; Particle reconstruction; Convolutional neural network; Volumetric particle image velocimetry
• ISSN: 2524-6992; 2524-6992
• DOI: 10.1186/s42774-021-00087-6

Three-dimensional particle reconstruction with limited two-dimensional projections is an under-determined inverse problem that the exact solution is often difficult to be obtained. In general, approximate solutions can be obtained by iterative optimization methods. In the current work, a practical particle reconstruction method based on a convolutional neural network (CNN) with geometry-informed features is proposed. The proposed technique can refine the particle reconstruction from a very coarse initial guess of particle distribution that is generated by any traditional algebraic reconstruction technique (ART) based methods. Compared with available ART-based algorithms, the novel technique makes significant improvements in terms of reconstruction quality, robustness to noise, and at least an order of magnitude faster in the offline stage.