Non-matching Nan indexes in u_original and v_original

[Sebastien Galas]

Hi,

I tried to unckeck every pre-processing and post-processing steps so the only thing PIVlab does is FFT PIV passes on raw data. I end up finding Nans in u_original and v_original and these Nan are not at the same index when looking at the same image pair.

I do not understand why I have Nan if I only do the FFT PIV step, and most importantly I do not understand why they are not in the same spot for u and v.

Can you please help me with what the analyse section in PIVlab exactly does?

[William Thielicke (Shrediquette)]

Hi, if you cross-correlate two black sub images, then the result is NaN. Also, the sub pixel estimator might fail if there is too many completely black or white pixels. And X and y direction are independent from one another, so NaN might appear at different indices.

[Sebastien Galas]

thanks for the help!