How to calibrate

[omata julius]

Dear Mr williams. Is it possible to calibrate a random data set i found online? If yes how can it be when i dont know the real distance between any thing in the image

[William]

Hi, then it cannot be done. You need to know a distance to have a conversion factor.

[omata julius]

thankyou Dr Williams. I saw this geo reference link https://www.sciencebase.gov/catalog/item/5fb7b60dd34eb413d5e14811 . I dont know if this makes any meaning for calibration since i do not know any known distance

[omata julius]

Can this geo referenced pix and information be used as an alternative to calibrate?

[William]

Yes, probably you can measure some distance e.g. using google earth. But both your reference image and the PIV images need to have the same resolution. Then you additionally need the delta t between the PIV images.

[omata julius]

Hello Dr Williams, Thankyou for your replies.

I hope this message finds you well. I noticed from the study that the scale used was 0.1 meters/pixel and the frame rate was 25 frames per second. Based on this, I applied a scaling factor of 0.04 for my calibration. However, I’ve encountered several issues:

1. When I apply the Ensemble Multipass FFT, it only processes a single image instead of running across the full image sequence.

2. I’m unable to compute the mean velocity—the option appears inactive or yields no result.

3. Any plot I attempt to generate (e.g., vector or contour plots) results in a dark or invalid output.

[William]

Dear Julius, ensemble correlation uses all frames of a session to generate one velocity map. This is how ensemble correlation works by principle. So this is already a mean velocity.

It is surprising that there is nothing displayed in the contour plots. Did you maybe disable transparency? What is your matlab version? I could recreate this issue with version 3.10 and matlab 2025a.

[omata julius]

Dear Dr. William,

Thank you for your reply. I’m trying to replicate Legleiter’s results using their reported scale of 0.1 m/pixel. I’m using MATLAB R2024a and PIVlab 3.09.

Unfortunately, I’m not getting any plots, and the velocity values seem completely off. Could you kindly guide me on how to set things up correctly? If possible, could you try to recreate the issue on your end?

[omata julius]

Also PIV lab is asking for my real distance in mm yet giving me a mean velocity in m/s. could this be why my mean velocity is so high by a 1000?

[omata julius]

Good Morning Dr williams. I am still waiting for your reply

[thielickeoptolution]

You have to enter the distance in mm. Why is your time step 0.04 ms? Was the data recorded with 25000 fps...?

[omata julius]

Good day Dr. I had to convert to seconds from ms. I actually used 40ms since its data was recorded in 25fps. but it still wasnt working out. No plot is showing and i cant derive any meaningful information from it.

[William]

Then maybe upload the session file and the image files here, so I can check why no plot is showing.