Sensitivity Analysis of Event Based Algorithms for Velocimetry

Building on recent validation studies, the present work performs a systematic sensitivity analysis of a weighted event-based tracking algorithm applied to two configurations. The first is freestream and post stall flow around an SD7003 wing in a water tunnel. The second is a freestream-cavity arrangement that combines fast outer flow with slow, highly sheared recirculation. The study quantifies how event camera bias settings, time segment duration, and internal tracking thresholds influence event density, computational cost, and velocity accuracy. Results show that high pass filtering and data volume primarily control computation time, whereas time segmentation and the minimum ellipse diameter have a stronger impact on the recovered mean velocity. In the slow cavity, temporal thresholds that govern inter event gaps and time linearity become the dominant factors. The analysis yields practical guidelines for choosing camera biases and tracking thresholds so that reliable velocity estimates can be obtained in both fast and slow regions of a flow while maintaining manageable data rates and processing times.