The evaluation of a crosscorrelation pattern recognition technique for flow visualization

Abstract

A method based on crosscorrelation of synthesized digital image pairs has been used for determination of velocity vector fields associated with fluid flow about a cylinder. A simple model has been used to test the validity of this image-processing technique for replacement of the time-consuming manual evaluation of time-sequenced photographs used for flow visualization. A time-sequenced digital image pair was synthesized by using the standard equations of motion of fluid flow about a cylinder. Both optimum particle density and time delay for the best performance of the algorithm were found. © 1990.