Share Email Print
cover

Proceedings Paper

An optical flow-based method for velocity field of fluid flow estimation
Author(s): Grzegorz Głomb; Grzegorz Świrniak; Janusz Mroczka
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

The aim of this paper is to present a method for estimating flow-velocity vector fields using the Lucas-Kanade algorithm. The optical flow measurements are based on the Particle Image Velocimetry (PIV) technique, which is commonly used in fluid mechanics laboratories in both research institutes and industry. Common approaches for an optical characterization of velocity fields base on computation of partial derivatives of the image intensity using finite differences. Nevertheless, the accuracy of velocity field computations is low due to the fact that an exact estimation of spatial derivatives is very difficult in presence of rapid intensity changes in the PIV images, caused by particles having small diameters. The method discussed in this paper solves this problem by interpolating the PIV images using Gaussian radial basis functions. This provides a significant improvement in the accuracy of the velocity estimation but, more importantly, allows for the evaluation of the derivatives in intermediate points between pixels. Numerical analysis proves that the method is able to estimate even a separate vector for each particle with a 5× 5 px2 window, whereas a classical correlation-based method needs at least 4 particle images. With the use of a specialized multi-step hybrid approach to data analysis the method improves the estimation of the particle displacement far above 1 px.

Paper Details

Date Published: 26 June 2017
PDF: 11 pages
Proc. SPIE 10329, Optical Measurement Systems for Industrial Inspection X, 103293P (26 June 2017); doi: 10.1117/12.2270086
Show Author Affiliations
Grzegorz Głomb, Wroclaw Univ. of Technology (Poland)
Grzegorz Świrniak, Wroclaw Univ. of Technology (Poland)
Janusz Mroczka, Wroclaw Univ. of Technology (Poland)


Published in SPIE Proceedings Vol. 10329:
Optical Measurement Systems for Industrial Inspection X
Peter Lehmann; Wolfgang Osten; Armando Albertazzi Gonçalves, Editor(s)

© SPIE. Terms of Use
Back to Top