Affiliations: Department of Mechanical Engineering, Saitama
University, Urawa, Saitama, 338-8570 Japan | Fuji Heavy Industries Ltd., Ohta, Gunma, 373
Japan
Abstract: Speckle Velocimetry or high-image density PIV gives a velocity
vector map of a two-dimensional flow field by point-by-point spatial frequency
analysis of local pattern at an interrogation spot in a double- or a
multiple-exposed image of laser speckle or pseudo-speckle pattern generated by
pulsed laser-light-sheet illuminations of a plane in the flow densely seeded
with fine particles. The whole field spatial frequency analysis of the double-
or multiple-exposed PIV image gives more useful information of the flow field.
Optical Fourier transform is a conventional technique not only for the local
spatial frequency analysis but for the whole field analysis. Filtering of
spatial frequency is one of the typical techniques for the latter which can
reconstruct a velocity contour or a component velocity contour map of the flow.
Fundamentally this technique is a simple and efficient analogue method to get
more information in the velocity field analysis of the flow than a digital
image processing, but in practical applications the measurement is restricted
to simple flows because of optical noise and low spatial resolution. In order
to improve the technique, the fundamental characteristics of the filtering and
the noise yielded in the filtering process must be investigated. Meanwhile,
wavelet analysis can also be applied to the whole field spatial frequency
analysis of PIV image. In this paper the filtering technique is examined by
numerical convolution integral, and the results obtained are compared with ones
obtained by the wavelet analysis.