Event-based 3D-Lagrangian Particle Tracking using Different Processing Strategies

Kurzfassung

For image-based measurements in turbulent flow, the displacement of the particles reaches values of several 100.000 pixel per second, requiring fast cameras for time-resolved observation. As turbulence is a three-dimensional phenomenon, multiple cameras may be used to resolve the volumetric particle motion. We describe several approaches to retrieve time-resolved particle tracking data from event-based particle imagery acquired by multiple synchronized event-cameras in a multi-view arrangement. Statistical convergence requires measurement times of up to several minutes which would incur very large data amounts of O(10 Terabyte) when using multiple high-speed framing cameras. The costs of such systems along with powerful laser can be in the order of half a million Euro. As a possible alternative, event-cameras along with a cw-laser lowers the cost by two orders of magnitude. The data for a similar measurement time also reduces to O(100 Gigabyte). However, the reduced data stream imposes a variety of data processing challenges. In order to assess the pro’s and con’s of different event data processing pipelines, experiments capturing the near wall region of turbulent boundary layers were performed using either continuous or pulsed illumination of the measurement volume. While the different processing pipelines all yield time-resolved 3d trajectories, the underlying processing schemes each have unique benefits and challenges which are discussed in the contribution. The application of Shake-The-Box particle tracking pseudo-frame sequences obtained by pulsed illumination of the particle field proved to be successful, given proper image pre-processing. Gaussian smoothing of the images allows for the application of established peak detection and position optimizations schemes, as used within STB. The techniques should be tested on more datasets and measurement configurations to investigate tracking stability under varying experimental conditions.