myimagelib.pivLib.PIV

myimagelib.pivLib.PIV(I0, I1, winsize, overlap=None)

Standard PIV, consisting of replacing outliers, validate signal to noise ratio, and a smoothing with median filter of kernal shape (3, 3).

Parameters:

  • I0 (2D numpy.array) – The first image

  • I1 (2D numpy.array) – The second iamge

  • winsize (int) – interrogation window size

  • overlap (int) – distance between two windows, usually set to half of window size. By default, it is set to half of the window size.

Returns:

x, y, u, v – 2D matrices of PIV data. The shape of x, y, u, v are the same. The shape is (row, col), where row = (I0.shape[0] - winsize) // overlap + 1 and col = (I0.shape[1] - winsize) // overlap + 1.

Return type:

4-tuple of 2D numpy.array

>>> from myimagelib import PIV
>>> PIV(I0, I1, 64)
>>> PIV(I0, I1, 64, 32)

Edit

  • Nov 17, 2022 – (i) Turn off median filter. If needed, do it on the outcome, outside this function. (ii) Update the use of openpiv.pyprocess.get_coordinates().

  • Dec 06, 2022 – Update syntax per the changes in the openpiv module. Copy from tutorial.

  • Jan 12, 2023 – Change sig2noise threshold to 1.

  • Mar 02, 2025 – (i) Remove dt parameter. dt is set to 1.0, so that the unit of velocity is always pixel/frame; (ii) set default value of overlap to half of the window size. Make overlap optional.