Time-series image velocimetry based on time-delay estimation: a proof-of-concept study

Abstract In fluid image velocimetry, velocity is conventionally estimated from the inter-frame displacement of image features (e.g., artificial tracer particles). However, this paradigm breaks down when the displacement exceeds the field of view, leading to observation failure. To address this limitation, this study proposes a time-series image velocimetry framework integrating the underlying principles of the pixel intensity cross-correlation and space-time image velocimetry methods. By accounting for collective flow characteristics and space-time transformation relationships within a Eulerian framework, we formulate the time-series image velocimetry method. Furthermore, the generalized cross-correlation method is employed to estimate the time-delay. The proposed framework is validated using time-series images under varying flow conditions, particle concentrations, and noise levels. The results demonstrate that the proposed method effectively mitigates observation failure. This work provides a reframing and integration of existing temporal-velocimetry concepts to address observation failure.