SMuK 2021 – wissenschaftliches Programm

Bereiche | Tage | Auswahl | Suche | Aktualisierungen | Downloads | Hilfe

P: Fachverband Plasmaphysik

P 17: Poster II

P 17.3: Poster

Freitag, 3. September 2021, 14:00–16:00, P

Fast 3D particle position reconstruction using a neural network — •Michael Himpel and André Melzer — Institut für Physik, Universität Greifswald

We present an algorithm to reconstruct the three-dimensional positions of a particle cloud using a convolutional neural network. The approach is found to be very fast (thousands of particles in less than one second) despite its relatively high accuracy. It is special to this algorithm that the computing workload is separated into two parts: The energy- and time comsuming training, followed by the framewise non-intense and fast reconstruction. This makes this algorighm especially suitable for remote applications. The algorithm is applied to synthetic as well as experimental data from parabolic flights.