Modeling and Information Processing
Processing of PIV data
Using the GPU to speed-up PIV processing
1- Particle Image Velocimetry
Particle Image Velocimetry (PIV) is an experimental technique that enables the measurement of fluid velocity fields through the seeding of fluid with small particles that are imaged at a high rate with highspeed cameras coupled with a laser flash, cf.
Brossard et al 2009.
The following images provide a sketch of a PIV experiment in fluid dynamics, of a PIV image obtained by a PIV system and also of the vizualization of the velocity field estimated by FOLKI-PIV.
2-FOLKI-PIV
The teams of DAFE and DTIM have jointly developped the "FOLKI-PIV" [1] software that offers a high-speed/high-accuracy processing of image sequences generated by PIV in a wind tunnel.
This software embeds a FOLKI algorithm core and a suite of pre-processing and post-processing tools that increase the robustness of FOLKI with respect to PIV image specific features (long range illuination variation, walls...) and provide the usual visualization modes of the PIV community.
An illustration of the effectiveness of variation of illumination correction is proposed herafter. Both PIV images, corrected and uncorrected appear in a snapshot of FOLKI-PIV's GUI.
FOLKI-PIV also processes
stéréo-PIV [2] images in order to obtain the third component of the velocity on the laser sheet.
The following images provide an outlook of the FOLKI-PIV GUI for the case of stereo-PIV data from the 3
rd PIV challenge (Case E):
Left : Ground truth, Right : FOLKI-PIV estimate
Consult FOLKI-PIV
users manual for an outlook of FOLKI-PIV capabilities.
Executable binaries for a demonstration of FOLKI-PIV can be downloaded
here
3-Bibliography
[1] F. Champagnat, A. Plyer, G. Le Besnerais, B. Leclaire, S. Davoust et Y. Le Sant, “Fast and Accurate PIV computation using highly parallel iterative correlation maximization,” Experiments in Fluids, 2011. Preprint
[2] B. Leclaire, Y. Le Sant, S. Davoust, G. Le Besnerais, et F. Champagnat, “FOLKI-SPIV: a new, ultra-fast approach for Stereo PIV'' submitted to Experiments in Fluids, 2010. PDF
