PIVPROC is a computer program for processing data in particle-image velocimetry (PIV), which is a method of determining a flow velocity field from images of small seed particles that are entrained in the flow and that are illuminated by laser pulses at known intervals of time. PIVPROC creates an interactive computing environment for displaying and processing PIV image data. This environment includes a graphical user interface that provides user-friendly access to the image-data-processing capabilities of the program.
In PIV, a charge-coupled-device (CCD) camera records images that show the positions of the illuminated particles at two or more instants of time; then the image data are pro-cessed to extract velocities from the apparent displacements of the particles during the intervals between exposures. The processing involves one or more of three types of data-reduction techniques: autocorrelation, cross-correlation, and particle tracking. Autocorrelation is used to process double-exposure images, whereas cross-correlation and particle-tracking techniques are applied to pairs of single-exposure images.
In correlation processing, an image frame is divided into small subregions, each containing particle images. An auto- or cross-correlation operation is performed in each subregion, wherein the average displacement of the particles results in a displacement peak on a correlation plane. From the location of the displacement peak on the correlation plane and the time between laser pulses, the velocity in the subregion during that time can be computed. By thus processing the image over a regular grid of small subregions, one generates a velocity-vector map.
In particle tracking, displacements of individual particles are identified and used to compute velocities. In a combined correlation/particle-tracking operation, a correlation velocity-vector map is computed, then used as a guide for particle tracking.
The graphical work environment created by PIVPROC helps the user to perform autocorrelation, cross-correlation, and particle tracking operations on PIV image data. The raw PIV image data can be loaded and displayed on the computer screen. The image gain and threshold level can be adjusted by use of dialog boxes. The correlation processing settings are also displayed in dialog boxes. Subregions wherein correlation processing is in progress are displayed in real time, along with the output correlation plane.
PIVPROC employs fuzzy logic for validating detections of correlation peaks and for determining correct particle pairings in particle-tracking operations; fuzzy logic enables the implementation of data-reduction algorithms that mimic or surpass the ability of a human operator to identify the correct particle pairings in the image data. PIVPROC supports the combined use of cross-correlation and particle tracking to obtain high-quality velocity data over a wide range of particle-seeding densities.
The velocity-vector maps generated by processing image data can be displayed, edited manually by use of the computer mouse, printed, and written to files. The data can also be interpolated; the program includes interpolation algorithms that enable the user to transform the spatially randomly sampled data from a particle-tracking operation onto a uniform grid of velocity vectors; the use of a uniform grid facilitates comparison of a velocity field determined by PIV with the corresponding velocity field determined by computational fluid dynamics.
PIVPROC runs in the Windows 95, Windows 98, and Windows NT operating systems. All of the data-processing and image-manipulation routines in PIVPROC are written in FORTRAN. The Microsoft Windows Application Programming Interface (API) functions are used to generate and service the interactive user environment.
This work was done by Mark P. Wernet ofGlenn Research Center. LEW-16857