proPTV: A probability-based particle tracking velocimetry framework

Abstract

We present proPTV, a comprehensive framework for particle tracking velocimetry. The framework is an open-source software project and written in Python. It provides the user with all tools needed to process raw camera images of a particle-seeded flow in order to reconstruct the particle dynamics and to assimilate pressure fields. The advanced probabilistic tracking method enables accurate reconstruction of the most probable particle trajectories. Its performance is studied applying it to the numerical test case of turbulent Rayleigh-Bénard convection (Ra=1E10, Pr=6.9) in a cubic cell generated by direct numerical simulation. For the highest tracer particle density of about 0.125 ppp of this test case, 83% of the reconstructed tracks are correct. To check the performance also for experimental data, proPTV is additionally applied to particle measurements of turbulent Rayleigh-Bénard convection in a water-filled cell for similar Ra- and Pr-numbers as the numerical test case. Thereby, a tracer particle density of about 0.02 ppp is estimated. The obtained results are then compared with those obtained using LaVision's commercial particle tracking software DaVis (v10.2.1). Both frameworks provide velocity fields that have small deviations. However, the particle tracks generated by proPTV are on average 5 times longer than those generated with DaVis. proPTV including the numerical test case is available at: https://github.com/RobinBarta/proPTV.