Is there a comprehensive tutorial for openpiv python?

[Can Özcan]

Hello OpenPIV;

I have joined this group a while ago in hopes of using PIV in my phd thesis. For past ~2 years I was struggling to get an very basic PIV setup. Thanks to Alex Liberzon he referred some papers for simple PIV systems. Starting from there I have tried many things. I was able to get some image pairs (an example in attached zip file). I know it is not ultimate quality but I feel like I can use it to perform some PIV analysis. 

My personal preference is to use openPIV python to analyze these images. However I am having a hard time using openPIV due to lack of documentation. I have seen the following sources:

- https://openpiv.readthedocs.io/en/latest/index.html#

- https://nbviewer.jupyter.org/github/alexlib/openpiv-python/tree/master/openpiv/examples/notebooks/

But to be quite honest the lack of some explanation about the problems is preventing me moving forward as I want to. For example there is some scaling used in tutorials as "96.52". Where is this coming from? I assume it is based on some pixels/length measurement. 

I have tried using PIVLab and was able to generate some results easily. The tutorial in this page (https://pivlab.blogspot.com/p/blog-page_19.html) is very nice and comprehensive. But I prefer using python as it is open source and free for all. 

Can you please guide me if there is such tutorial for openPiv?

I am willing to put some effort into developing one if someone can give me a hand where/how to find relevant functionality of PIVLab in openPIV python version.

Best regards.

- 

Can Ozcan, M.Sc. in Mechanical Engineering

PhD. Cand. in Biomedical Engineering

AKRO Mühendislik Hizmetleri Araştırma ve Geliştirme Ticaret Limited Şti.

İzmir Teknopark İYTE Sit. A10 Bina

Gülbahçe Mah. Gülbahçe Cad. No: 1/40/31 

URLA/IZMIR/TURKEY - 35437

Ph: 90-232-503 05 18 | Fax: 90-232 503 05 18 | Cell: 90-532-693 2780

can....@akromuhendislik.com | www.akromuhendislik.com

   

[Peter Vennemann]

Dear Can,

you may try the Tk-GUI:

https://openpiv-tk-gui.readthedocs.io/en/latest/

All parameters are documented with tooltips.

Regards,

Peter

[alex.l...@gmail.com]

Dear Can, 

This is a great idea - we always need help with the documentation. 

Thanks to Peter (see below) the GUI version in Tk is better documented with tooltips. Nevertheless, I suggest to get your help and convert tooltips also to a proper tutorial, following the one of PIVLab or others. 

What we have so far is 

Main documentation would be always: 

https://openpiv.readthedocs.io/en/latest/ 

We shall write it and improve it. Instead of a tutorial it now points to some notebooks 

https://openpiv.readthedocs.io/en/latest/src/tutorial.html

So the next steps would be :

1) move these notebooks to the tutorial itself under the documentation. I already know how to do it using jupyter in sphinx (here's an example using just jupyter notebooks) 

2) copy some stuff from the old tutorial https://alexlib.github.io/openpiv-python/index.html 

3) copy things from the sporadic tutorials here and there, https://github.com/OpenPIV/openpiv_basics or http://www.openpiv.net/gettingstarted.html and from all the examples: https://github.com/OpenPIV/openpiv-python/tree/master/openpiv/examples/notebooks 

Please, join me on the github (write an issue in the https://github.com/OpenPIV/openpiv-python/issues) and we will start working on it. I strongly recommend to use Jupyter noteboks - this helps both to create a single set of documentation and also an ability for the user to click and open a cloud Binder to test it with the present installation. This will also create a lot of examples one could follow. 

Of course if you have another idea on how to move on with the documentation, don't hesitate to write here. 

[alex.l...@gmail.com]

Thanks Can for reminding us. I've checked now that we can build the documentation and extended with the option to use Jupyter notebooks in it. 

Looks much better. So we can now add more and more explanations using just Jupyter notebooks. https://openpiv.readthedocs.io/en/latest/src/piv_basics.html

[Can Özcan]

Thank you for the opportunity Alex. 

I have gone to github and saw you already created the issue. And probably made some updates as well. I will try to follow-up there to make some contribution to the documentation, along with my own education :)

I also would like to resend the following link to the group, which is a presentation given by you and Hadar Ben Gida, Roi Gurka. I found this very useful, but it is going thru Matlab. I highly recommend this video to the audience.

https://www.youtube.com/watch?v=ci98mLhYEeg&t=904s

I have one basic question at this point: Does OpenPIV python version has all the capabilities of the matlab version? I am not referring to the GUI but to the "engine" for "model" code base.

Best Regards...

--
You received this message because you are subscribed to the Google Groups "openpiv-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openpiv-user...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/openpiv-users/60792f07-0ee9-4bf5-8476-4491a6a50d01n%40googlegroups.com.

--

[Can Özcan]

Sorry, the answer of my question "I have one basic question at this point: Does OpenPIV python version has all the capabilities of the matlab version? I am not referring to the GUI but to the "engine" for "model" code base." is towards the end of the video I posted. 

Matlab version has more features and python does not have one-to-one match for those. But may have other features.

Best.

[alex.l...@gmail.com]

Thanks for recommending the webinar - it is indeed designed to be useful. Thanks to Roi Gurka and Hadar Ben Guida who're actually the main contributors to this webinar. 

To answer your question: 

1) Python version has windows deformation and 3D PIV algorithms that Matlab version does not have

2) Matlab has a GUI that was used from the beginning and never updated seriously. Python now has great contribution in Tk GUI https://github.com/OpenPIV/openpiv_tk_gui. There is also a Qt based one, but since I don't have to develop it further, it's stuck https://github.com/OpenPIV/piv-gui-python

3) Python as all the great service of Jupyter Notebooks, MyBinder and Google Colaboratory, as well as other cloud services, e.g. it can be serverless (not yet) https://github.com/OpenPIV/serverless_OpenPIV 

4) Python has high-end GPU extension running on the clusters - https://www.researchgate.net/publication/332239196_GPU_Accelerated_Open_Source_Particle_Image_Velocimetry_Software_for_High_Performance_Computing_Systems and https://github.com/OpenPIV/openpiv-python-gpu 

5) Python is easy to integrate into a webserver https://github.com/OpenPIV/openpiv-saas 

6) I probably forgot what else it has :)