Advanced optimization of correlation imaging velocimetry algorithms

Abstract

Significant developments aimed at improving the performance of digital particle imaging velocimetry (DPIV) type techniques are presented. The analysis applies directly to the correlation imaging velocity (CIV) technique but is applicable to all forms of imaging velocimetry measurement. These improvements revolve around a hierarchical processing scheme, and take into account the effects of the local fluid deformation calculated from successive passes. Crucial to the implementation of these techniques is the ability to transform the discrete image into a continuous function that can be deformed and resampled at arbitrary resolution. Alternative methods of sub-pixel positioning are presented along with a new diagnostic tool for evaluating the severity of peak-locking errors in actual laboratory data. The performance of the new algorithms is illustrated with examples of turbulent laboratory flows, and simulated images obtained from analytic flow solutions or direct numerical simulations (DNS).