Dealing with noisy data

[Michał Kalkowski]

Hi,

I'm currently trying to reconstruct an EBSD scan composed of many files and stitch it together. The data is very noisy, and I have not succeeded in finding good reconstruction parameters yet. The reconstruction doesn't need to be very fine-detailed and for my purposes, a considerable amount of averaging is allowed. However, the starting point is quite troublesome. When I import the .ang file, without any cleanup operations I get the IPF figure as attached. By comparison, if I just load the same file into MTEX, I receive a much better quality IPF graph, which would be a better starting point for segmentation in Dream.3D. I do not know MTEX and prefer not to learn it, but I apply no operations there just import the data.

I attach both images for comparison. They come from just loading the exact same .ang file.

D3d:

MTEX:

Does anyone have any suggestions on how to treat noisy data in Dream.3D?

Why there is such a difference between the two packages from just parsing the EDAX (.ang) data?

Thanks!

Michal

[Michael Jackson]

Would it be possible to get the .ang file? Usually there isn’t a difference between MTEX and DREAM3D aside from a slightly different “Shade” of the IPF colors. MTEX might be making an assumption about the data that DREAM3D is not making? Maybe the header portion of the .ang file is not correct and the Laue class isn’t being set correctly during import? If the .ang file was generated by something outside of EDAX software, did it save the Euler angles as Degrees instead of Radians? Ang files should be in radians already although there is a simple filter in DREAM3D to convert from degrees to radians if you want to give that a try: (Convert Angles to Degrees or Radians). Those are all of the obvious things that I can think of.

 

What version of DREAM.3D are you using?

 

--

Michael Jackson | Owner, President

      BlueQuartz Software

[e] mike.j...@bluequartz.net

[w] www.bluequartz.net

 

 

From: <dream3...@googlegroups.com> on behalf of Michał Kalkowski <kalko...@gmail.com>
Date: Friday, January 15, 2021 at 8:16 AM
To: dream3d-users <dream3...@googlegroups.com>
Subject: Dealing with noisy data

 

Hi,

 

I'm currently trying to reconstruct an EBSD scan composed of many files and stitch it together. The data is very noisy, and I have not succeeded in finding good reconstruction parameters yet. The reconstruction doesn't need to be very fine-detailed and for my purposes, a considerable amount of averaging is allowed. However, the starting point is quite troublesome. When I import the .ang file, without any cleanup operations I get the IPF figure as attached. By comparison, if I just load the same file into MTEX, I receive a much better quality IPF graph, which would be a better starting point for segmentation in Dream.3D. I do not know MTEX and prefer not to learn it, but I apply no operations there just import the data.

I attach both images for comparison. They come from just loading the exact same .ang file.

D3d:

MTEX:

Does anyone have any suggestions on how to treat noisy data in Dream.3D?

Why there is such a difference between the two packages from just parsing the EDAX (.ang) data?

 

Thanks!

Michal

 

--
You received this message because you are subscribed to the Google Groups "dream3d-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dream3d-user...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dream3d-users/aed37cb9-5561-4bda-8466-ccb1ef58ae9an%40googlegroups.com.