Imaging for SEM using Pycroscopy and an Acquisition Hardware

[p chx]

Hello,

I am very happy to discover an open-source project as Pycroscopy. I am not user at the moment, I am just interested by the project. I enjoy any kind of open-source instrumentation overall.

To follow I am looking for a way to replace the Acquisition System on an old Scanning Electron Microscope. I mean first of all, the 2D scanning coils control hardware, the signal acquisition tool after the SE detector and the Image Processing Software. All open-source.

It existing a couple of commercial solutions, which are black box system and relatively difficult to fix by yourself if something wrong.

My goal is to develop an open-source hardware for SEMs, and Pycrosopy could be an amazing solution to imaging from the row data.

In this prospect, I am wondering:

-1- As there is not a huge gap between STEM and SEM in the beam control, do you think that Pycrospopy could be an adapted solution to do the proper image processing used in SEM ? 

-2- For a such thing, the image rate should be enough significant (as an SEM work usually). Is Pycrocopy could be adapted for that ?

Thanks for any help!

Pierre

[Rama Vasudevan]

Hi Pierre,

Welcome to the pycroscopy forum! It is very exciting that you are developing an open source beam control for SEM. 

With regards to the questions, I see no issue with image processing of SEM data using pycroscopy. Much of our processing is rather data agnostic, in the sense that it doesn't matter what type of image/spectral data you have (there are exceptions to this, but the majority are not). Since you will open source the acquisition, it probably makes sense to write files directly in the original USID format that pycroscopy leverages. 

The question of processing at image rates is more tricky. We have some deep learning packages that are sister projects to pycroscopy, that could be useful for realtime image processing. Those are not incorporated into pycroscopy yet, but we do aim to do this in due course. Incidentally, we have weekly hacakthons (via Microsoft Teams) every Friday 3PM-5PM, USA Eastern daylight time. Please let me know if you would like to join. 

Kind Regards,

Rama