matlab - PIV Analysis, Interrogation Area of The Cross Correlation -


i'm running piv analysis on 2 consecutive images taken during experiment vector field. know, based on criteria have choose percentage of overlap between tow images cross-correlation process? 50%, 75%...? pivlab_gui tool designed matlab chooses 50% overlap default, allows changing it.
want know criteria based on can know how overlap best? vectors become less accurate, dependent.etc, increase/decrease overlap?
book "fluid mechanics measurements" not explain how choose overlap amount in cross-correlation process, , not find helpful online reference.
appreciated.

i suggest read on spectral estimation - equivalent cross correlation when segment data , average correlation estimates calculated each segment (the cross correlation inverse fourier transform of cross spectrum). there's book chapter on stuff here, may want find more complete resource if unclear on basics.

a short answer: increasing overlap increase frequency resolution of spectral estimate, , give more segments average over; estimate have lower variance. there diminishing statistical returns more increase overlap past 50%, while computational complexity continues rise (more segments = more calculations). hence people choose 50% , have done it.

it's important note don't more information using overlapping frames, increasing frequency resolution (or time lag resolution, correlation) - similar effect of zero-padding signal before taking fourier transform - , has statistical effects due way estimation of type works.


Comments

Popular posts from this blog

Perl - how to grep a block of text from a file -

delphi - How to remove all the grips on a coolbar if I have several coolbands? -

javascript - Animating array of divs; only the final element is modified -