wiki:Tutorial/DisplacementMeasuredByImageCorrelation

Version 9 (modified by vaillant1p, 9 years ago) (diff)

--

Tutorial / Displacement Measured by Image Correlation

This demo shows how to use image correlation to measure a displacement versus time (instead of velocity). Open the image UVMAT_DEMO06_displacement (accessible on http://servforge.legi.grenoble-inp.fr/pub/soft-uvmat/).

Looking at the images:

Open an image UVMAT_DEMO06_displacement. This is a target pattern used to measure the displacement of a plate vs time. Observe the motion by scanning the images.

Make a mask to remove the outer part in the PIV processing, like explained for tutorial 5, or introduce the mask provided .ref (selecting the upper bar button .. in uvmat)

Civ in mode 'displacement':

Open series/civ_series. import Config for ref file, .ref . The appropriate parameters are displayed:

  • mode displacement, with respect the origin index 1 (first image by default). Then the menu for choosing image pairs desappears, as the correlation for each image are done with respect to the reference one, taken as origin for the displacement.
  • Correlation box fairly large 31 corresponding to a coarse pattern, and considering that spatial resolution is not critical.
  • Large search 107 need to capture the max displacement
  • !CorrSmooth =2 corresponding top carse 'particles': large correlation curves. See !TestCiv1.
  • Mask selected to isolate the active area
  • Fairly large FieldSmooth =20 (spatial resolution not critical) and no 'deformation' needed for civ2
  • Civ2, same correlation box, smaller search zone as we start form the prediction of Civ2 (shift)

Observing the civ data:

velocity, vorticity, div

Time series:

Make a point, range=200, and run series/time series. Observe the curve displacement U and V versus time.

+ rotation + div

Attachments (2)

Download all attachments as: .zip