Posts

Please support PIVlab!

Please support PIVlab here: https://paypal.me/pools/c/8cj8E4slJx
I have submitted another paper on PIVlab in an open access journal. It describes recent additions to PIVlab, tests precision and robustness. But it is freaking expensive to publish (for a hobby project). I asked for a discount, we'll see if I get one, but it is still expensive. 
Let's be honest: I will personally benefit from this publication, as people will most likely cite it. And maybe even more people will use PIVlab for serious research, which makes me a bit proud.
So why should you contribute to a publication? Well, I think that the content might also help you (and maybe also the reviewers of your publication) to understand PIVlab better and trust in PIVlab even more. So feel free to contribute!
Futhermore, Matlab updates sometimes make PIVlab incompatible with new Matlab releases. Therefore, I need to renew my own Matlab license regularly to keep track changes, making sure that PIVlab will keep on running flawle…

PIVlab is now on GitHub!

Image
Since quite a while, Matlab File Exchange offers the opportunity to link to a GitHub repository. So I just created a PIVlab repository on GitHub, and changed the Matlab File Exchange site accordingly. I hope that this works out smoothly, and everyone benefits from this change!
Here is the link to GitHub: https://github.com/Shrediquette/PIVlab
And here is the modified File Exchange site: https://de.mathworks.com/matlabcentral/fileexchange/27659


PIVlab Update v2.36

Ensemble correlation: Performance upgrade and bug fix (with old matlab versions)Saving huge PIVlab sessions (> 2GB) now works more reliableEnhanced compatibility with old Matlab releasesUpdated uipickfiles to latest versionPIVlab more reliable remembers last folder and last import settings(PIVlab v2.35 contained a bug, so it was replaced with 2.36 instantly)

Vortex street analysis in 4k and 60 fps

Image
This is some data that I recorded quite a while ago, but I just discovered it on a backup HDD. The output data of PIVlab was edited and rendered in high resolution.
Click here to watch it in full resolution




PIVlab Update 2.31

PIVlab 2.31 is released:

New features:

Ensemble correlation with all the features of the existing algorithm (muti pass window deformation, repeated correlation, disabling self correlation, etc).Information here and here.
Fixes:

Quick acces toggle button status fixedBug in a visibility property fixedOpening Shortcut PDF throws a more appropriate error message when it fails

The benefit of ensemble correlation in PIVlab v2.30

Image
PIVlab v2.30 features 'multipass window deformation ensemble correlation'. This type of correlation has been introduced by Santiago et al. in 1998. It is especially helpful in micron-resolution particle image velocimetry (micro-PIV, µPIV), as it can deal with very low seeding densities (see below). Ensemble correlation PIV should be used if you want to analyze a steady flow that has a low seeding density. You need to record a large number of images to have the full benefit of ensemble correlation in PIVlab.



Ensemble correlation is now fully integrated in PIVlab v2.30:



And here, you can see how the quality increases with increasing number of images:


Soon: Ensemble correlation!

Image
I am currently working on the implementation of ensemble correlation.

This kind of correlation averages the correlation matrices of a frame series before searching for a peak. Using it makes sense, when there a only a few particles visible in the flow. This is often the case in micro-piv, and I hope that there will be many users that benefit from this feature.

The changes to the GUI are finished, and I have a simple 1-pass ensemble correlation running. But I want to have the full potential of multi-pass window deformation algorithms with ensemble correlation. Implementation therefore takes a bit more time as there are some changes that challenge my brain. Additionally, most of the programming currently happens single handed because of my third daughter...


PIVlab update 2.2

Image
New features / enhancements:
New and enhanced correlation algorithm (linear correlation instead of circular correlation). Details here.Correlation quality dropdown menu:Normal: Circular correlation, linear window deformationBetter: Circular correlation, spline window deformationHigh: Linear correlation, spline window deformationExtreme: Linear correlation, spline window deformation, repeated correlationSuggestion for selecting an appropriate interrogation area size. Works by trying to count the amount of texture / particles and calculating the maximum displacement. PIVlab will try to suggest settings so that every window contains sufficient texture / particles and the displacement is not too large.Quick acces toolbar for the most important PIVlab functions. Will not be displayed if there is not enough space in PIVlab's program window.Automatic histogram limiting and stretching works now seperately for both images of a frame. This accounts better for unevenly illuminated imagesSynth…

Evaluation of the new PIVlab (v2.2) settings

Image
Recently, an interesting discussion, initiated by Theo, popped up in the PIVlab forum. Alex, the developer from OpenPIV also joined and gave valuable comments. The discussion was about whether circular correlation or linear correlation would be more suitable for processing PIV images (read the details in the link given above if you're interested).
PIVlab uses circular correlation since the beginning. Now, I also tried linear correlation, and the results were pretty good: Linear correlation decreases bias and random (RMS) error, and it also enhances the robustness (less wrong vectors in bad image data). However, the processing time increases by a factor of 2.5 too.
As there are too many settings in PIVlab (window deformation interpolator, repeated correlation, correlation type) that are difficult to understand for a standard user, I decided to remove these settings.
In the latest release (coming in a few days), there is instead a drop-down box that looks like this:


The user can sel…

My talk on the MATLAB EXPO 2019 in Munich (July, 2nd)

Image
I will be giving a talk on PIVlab on the "MATLAB EXPO 2019 Deutschland 2. Juli | München".
The link to the agenda is here: https://www.matlabexpo.com/de/2019/agenda.html
Maybe this would be a nice chance to meet...?


Update PIVlab 2.02

Image
Changes:New derivative: Vector direction (use HSV colormap for best display performance)Save and load PIVlab masks, also load masks from previously saved session filesMask transparency can be set, can also be used to make the mask invisible (100% transparent)Tool tips display help and information for all buttonsKeyboard shortcuts enables the user to work more efficientlyList with all keyboard shortcuts can be accessed from the help menuFixed a zoom bugWhen smoothing data, some extra information on what actually happens is now displayed in Matlabs command windowLoad external mask(s) improvedRemoved some unnecessary data in *.mat file export

Donations complete - thank you!

Image
As my personal Matlab license was pretty old, I was asking for donations for renewing my license during the last 365 days. It is important that I have a recent Matlab version to keep PIVlab alive.

And today, the donations reached 99.37 EUR, and I just renewed my Matlab license :-D Thank you! The new Matlab release even makes PIVlab 10.24 % faster!



Update PIVlab 2.01

Image
New features:
The user can now load time-resolved masks instead of single masksThe user can now select the color of the color legend labelThe PIV settings are now checked automatically, and a warning is given if unusual parameters were selected 
Bug fixes: Loading session files from older PIVlab versions is now possible (forum post)Fixed bug where the tools panel showed slightly incorrect data sometimes (forum post)Fixed bug where a warning message was displayed too often (forum post)




Major update: PIVlab 2.0

Image
In the past, it happened quite often that the user interface of PIVlab was not displayed correctly on Mac computers. So I decided to completely reprogram the user interface "by hand" (that means without using Matlabs built-in GUI IDE called "GUIDE"). It was quite some work, but I believe PIVlab is now future-proof and easier to maintain. GUIDE was crashing pretty often in the last months, making enhancements very difficult.



So these are the changes:GUI rewritten from scratch, should now be displayed correctly on all operating systemsAdded prefernces in the menu: You can change the panel width of the UI if the UI elements are not scaled correctly.Changed "Load external mask": You can now make masks outside of PIVlab and save them as black-and-white TIFs. When you import these masks, then white will become a masked area and black is non-masked area.Added TECPLOT file format for export.Background subtraction GUI only worked with grayscale images --> fixed…

Order a book about PIVlab!

Image
I still have a number of my dissertation books, it contains an extensive chapter on PIVlab. You can order a copy for 3 Euros + shipping (the money will be used to add to my Paypal money pool which will one day be used to renew my Matlab license).



The beauty of PIV cross-correlation!

Image
Do you like cross-correlation? You should when you visited this website...


PIVlab Update v1.50

Image
This update brings some new features and performance enhancements:

'Repeated correlation' I added the option 'repeated correlation'. I got the idea while I was working with another software (PIVview). 'Repeated correlation' can enhance your data yield when you are working with very noisy or generally bad PIV images. It does the cross-correlation not only once per vector, but 5 times in total. Each time, it slightly shifts the interrogation windows (the distance depends on the size of your interrogation area, but it's about 5-10 pixels). This happens during each pass in a multipass analysis. In the end, it multiplies the 5 resulting matrices, which results in less noise an a very clear peak. But it also takes about 3 times longer to calculate...

Background elimination The next big addition is the background elimination GUI. If you have recorded a longer time series of image data, it makes sometimes sense to calculate the average of all images, and subtract …