Measurement of time resolved velocities with large accelerations is challenging because the optimal capture rate and pixel resolution changes with velocity. It is known for velocity measurements that high temporal resolution and low pixel resolution increases the velocity uncertainty. This makes selecting acceptable camera settings unintuitive and can result in highly uncertain measurements. For experimental conditions with slow velocities (< 10 m/s) where high temporal resolution is required (because of rapid acceleration) there arises a need for exponentially increasing pixel resolution to minimize experimental uncertainty which is often impossible to achieve experimentally. Desired measurements for early flame propagation have velocities which span a wide range of velocity which can be greater than 10 m/s during ignition and can drop to under 1 m/s depending on the pressure. This rapid velocity change all usually occurs within a millisecond timeframe.
Typical camera-based velocity measurement usually observes either fast- or slow-moving objects with either an average velocity or a velocity at a single time. The goal of this work is to accurately measure such a rapidly changing experimental condition using camera-based measurement and understand the affect various processing methods have on the result. A practical method is presented here to quantify the noise and observe any induced errors from improper processing where measurable physical analogs are used to represent future experimental conditions. These experimental analogs are in the form of rotating disks which have known radial and velocity profiles that will enable the assessment of experimental parameters and postprocessing techniques. Parameters considered include pixel resolution, framerate, and smoothing techniques such as moving average, Whittaker, and Savitzky-Golay filters.
