Noise Robust Particle Event Velocimetry with A Kalman Filter-Based Tracking

Abstract

The necessity of understanding and analyzing flow field characteristics has prompted researchers to strive for more precise tools and systems to enhance the accuracy of flow velocity measurements. In this work, we introduce Particle Event Velocimetry (PEV), a novel approach to particle velocimetry that leverages an emerging imaging modality known as an event-based camera. We propose two variants of PEV: the Particle Event Velocimetry (PEV) and the causal Kalman Filter-Based Particle Event Velocimetry (KF-PEV) systems, which both are noise-robust particle-level estimates of the flow field based on the analysis of individual seed particles observed and tracked using the event-based camera. Compared to conventional frame-based (FB) Particle Image/Tracking Velocimetry (PIV/PTV) and event-based (EB) particle velocimetry methods, the proposed PEV’s and KF-PEV’s flow field achieves higher spatial resolution limited only by the density of the seeded particles while maintaining experimental simplicity. In addition, we also developed an event-based camera simulator for synthetic data with a ground-truth motion field to objectively benchmark the proposed PEV against other FB and EB methods and analyzed these competing methodologies’ ability to reconstruct sharp motion boundaries and field curvatures. Furthermore, both systems have been tested in real-world experimental environment at Water Tunnel Lab at the university of Dayton to validate its performance using real experiment of well known flow.