The applicability of large-diameter dense particles

[Cai Tian]

Hello Dr. William Thielicke and everyone:

My name is Tao, a student in HKUST. I am relatively new to using PIVlab for analysis so I apologize if this question seems rather simple. 

I am trying to calculate the vorticity of large dense particles ( diameter is 10mm, around 74pixels ) shearing within a ring cell as follows(only focus on the glass beads in the red box, around 724*723 pixels):

During the image pre-processing,  I use the "Enable CLAHE" and the chosen pixel is 128px, as follows:

During the analysis, I used the Multi-pass FFT window deformation method and chose three passes with window sizes of 256, 128, and 100 pixels, respectively, as follows:

I found the results are very sensitive to the setting mentioned about pre-processing and analysis, which makes me a little unconfident about my results.  So my question is: Is the PIV analysis suitable for my situation where the particles are very large and very dense(Diameter 74px, analysis area approximately 10d * 10d)? If possible, should I make modifications to my video recordings in those areas (such as darkening the glass beads to reduce reflection) or to the PIVlab processing (such as subtracting the minimum value or the average value during the preprocessing stage. The image showing the subtraction of the minimum value and the corresponding results are as follows) ?

Looking forward to your reply very much, and thanks again!

[William]

Hi, I think I would need raw, unedited input data to make a qualified suggestion.

[Tommy Yiu]

Thanks for your reply. I don't know why my reply was deleted, maybe because I add a sharing link?

I have updated the corresponding files in the ONEDRIVE( This time I delete the link here ), including the setting, extraction, and screenshots. 

The screenshots are as follows:

1. Input first 120 frames.

2.ROI setting.

3.Pre-processing.

4.Analysis.

5.Calibration

[Tommy Yiu]

Thanks for your kind reply.

The video has been shared in this OneDrive PIVTEST  . The corresponding files are including: movies, settings, mat files, extraction of mean values, and corresponding screenshots.

Details are:

1.To save time, only the first 120 frames have been selected. 

2. ROI settings.

3. Pre-processing.

4. Analysis settings.

5. Calibrations(diameter of glass bead is 10mm(around 62px in the movies)).

在2026年1月31日星期六 UTC+8 05:09:58<William> 写道：

[Tommy Yiu]

https://hkustconnect-my.sharepoint.com/:f:/g/personal/tyaoab_connect_ust_hk/IgBFKbh61hnKT6dc6DTmMzHHAT6BOeT9ECnKajPEJhRgRXc?e=vE4gye

在2026年1月31日星期六 UTC+8 05:09:58<William> 写道：

[Tommy Yiu]

Thanks for your reply. 

I have updated the corresponding files in the ONEDRIVE( PIVTEST  ), including the setting, extraction, and screenshots. 

The diameter of glass bead is 10mm, around 62px in the movies.

Key screenshots are as follows:

1. Input first 120 frames.

2. ROI setting.

3. Pre-processing setting.

4. Analysis stetting.

5. Calibration.

在2026年1月31日星期六 UTC+8 05:09:58<William> 写道：

[Tommy Yiu]

The link: https://hkustconnect-my.sharepoint.com/:f:/g/personal/tyaoab_connect_ust_hk/IgBFKbh61hnKT6dc6DTmMzHHAT6BOeT9ECnKajPEJhRgRXc?e=vE4gye

[William]

Hi Tao,

 I think PIV would work with these particles if you reduce the displacement. Currently, it is way too large (60 pixels in frame 1). The displacement should be limited to something around 5 - 10 pixels by increasing the framerate. If you do not have higher frame rate available, it could be feasible to use a camera with trigger input and frame-straddling.

Furthermore, the motion blur is excessive and must be reduced (more light, shorter shutter). You should not use video files but separate image files for better performance. I would also recommend to subtract the min intensity background image in PIVlab to reduce the correlation of scratches in the window. Was this captured with a mobile phone? I think you cannot expect to get good results with non-suitable camera equipment, sorry...