For the present analysis, multi-pass interrogation technique has been applied. In the first evaluation, the images were sub-divided into pixel non-overlapping interrogation images. The corresponding sub-images in the translated image and the reference image were analyzed by computation of the cross-correlation function. Since there was no overlap between adjacent interrogation images, each pair yielded 320 statistically independent displacement vectors. In the second evaluation, a pixel interrogation window was used with an overlap of 50%. The corresponding resolution is 0.5 mm and 5120 vectors were recovered from one pair of images. All measured integer displacements that deviated more than one pixel from the expected displacement were considered as spurious vectors. They were subsequently discarded from the data set. Such a strict test for spurious data can only be done if one has a priori knowledge of the displacement field.

Dynamic velocity and spatial range

Dynamic spatial range is related to spatial resolution and dynamic velocity range is related to the fundamental velocity resolution and hence, the accuracy of a PIV. Dynamic velocity range (DVR) specifies the range of velocity over which measurements can be made. It is the ratio of the maximum velocity to the minimum resolvable velocity, or equivalently the RMS error in the velocity measurement, i.e.

where is the image magnification and is the maximum time interval used for the experiments. The RMS error of the displacement field on the pixel plane ( ) generally lies between 1-10% and so

where represents the resolution of the recording medium that is taken to be equivalent to the pixel size, and is the diameter of the particle image prior to being recorded on the pixel plane. Assuming that the particle image is diffraction limited and its image intensity is Gaussian, the diameter of the diffracted image of the particle is expressed as: