An improved method for enhancing the resolution of conventional double-exposure single-frame particle image velocimetry

Abstract

An improved hybrid particle image velocimetry (PIV)/particle tracking velocimetry technique is presented. The method uses a conventional autocorrelation for a predictor step followed by local cross-correlations of individual particles to obtain individual particle pair displacements within a double-exposed single-frame image. Simulated particle fields are used to demonstrate the improved tracking success (better than 90%, even in very dense particle fields) and accuracy of the technique as compared to an existing single-frame hybrid method. In addition, the number of erroneous measured pairs is reduced. The improved success is attributed to the ability of the algorithm to differentiate between overlapped particles and the criteria for correctly identifying overlapped particles are presented. As a check on the effect on derived quantities, computed values before and after resolution enhancement from a real image with constant vorticity and dilatation are presented. Application of this method to a turbulent non-premixed flame shows about a factor of six increase in the resolution over conventional PIV.