Abstract
The particle imaging velocimetry (PIV) based on the pattern correlation method has an inherent problem of erroneous velocity vectors due to erroneous correlation. Two types of images were examined; one was the image of air flow visualized by a smoke tracer method and the other was the image of absolutely static water taken by a rotating video camera. The gray level difference method was also applied and the results were compared with those of the pattern correlation method. The probability of erroneous velocity vectors is strongly affected by the correlation matrix size as well as by the type of image.
