Super-time-resolution particle tracking velocimetry via the fusion of event- and frame- based cameras

Abstract This paper reports a new super-time-resolution particle tracking velocimetry (PTV) technique that uses a low-cost hardware fusion strategy comprising high-frequency event- and low-frequency frame-based cameras and a data fusion strategy involving event- and frame- based images. This novel PTV technique can enable long-time imaging and accurate particle-center position detection for particle tracking by using the position determined from grayscale frame-based images to correct that determined from non-grayscale event-based images. The effectiveness and accuracy of the fusion system and super-time-resolution PTV technique were examined via experimental measurement of jet flow at a Reynolds number of 9021. The results indicate that a long-time data sampling at a frequency of 2000 FPS was realized, allowing monitoring of the continuous time-series flow behaviors in an experimental system with a small data storage requirement. The super-time-resolution PTV technique realized a time resolution enhancement of at least 16 times compared with the frame-based camera, capturing finer particle motion details and reconstructing the time-resolved nonlinear motion process of the tracer particles compared to the low temporal resolution particle tracking. The fusion system improved the accuracy of particle tracking via position correction of low-frequency frame-based images, thereby reducing the average particle-center error to 0.4 pixels compared with the event-based camera only. Finally, long-duration data sampling was conducted in a circular impinging jet experiment (Re = 6766) to obtain time-averaged flow fields, thereby validating the measurement accuracy and long-term reliability of the super-time-resolution PTV technique with pulsed light illumination. Overall, the fusion system was shown to enable accurate long-time tracking of particles at a high spatiotemporal resolution, representing a new measurement technique in fluid mechanics and a novel form of data fusion for optical imaging of flow fields.