CORRECTION METHOD OF ERRONEOUS VECTORS IN PIV

Abstract

A new method that can detect and correct erroneous velocity vectors unexpectedly appeared in particle imaging velocimetry (PIV) is proposed. Two or three sets of four divergence values independently defined at a measuring point are used to examine local continuity of flow. The fundamental idea of the method is that some of these absolute divergence values must become greater than a certain threshold level if the local continuity is destroyed by erroneous vectors. Two criteria for the error detection are proposed and their efficiencies are examined by the simulation of uniform vector fields containing specified number of erroneous vectors. They are found to be satisfactory if the ratio of erroneous vectors is less than 0.5. It is also demonstrated that excellent performances are achieved when the presented method is applied to an experimental data of a complicated open-channel flow and a remote sensing data of flood flow.