Denoising EBSD

[saeadi]

Hello ,
My EBSD data shows the image below. None of the above commands make any difference to remove noise. My MTEX ​​version is 5.11.2. What am I doing wrong? How can I fix it?
Thanks a lot for your help.

[zmichels]

Can you share the dataset? Or more details? It's not fully clear to me from the image what's going on.

[saeadi]

Of course. The data was imported in ctf format and according to the following codes it gave me the above microstructure image. To reduce its noise I used the mentioned codes like mean and Kuwahara filters but no change was observed in the image. Not even smoothing of the grains is done!

[grains,ebsd.grainId] = calcGrains(ebsd,'angle',10*degree,'minPixel',5);
grains = smooth(grains,5);
ipfKey = ipfColorKey(ebsd.CS.properGroup);
plot(ebsd,ipfKey.orientation2color(ebsd.orientations))

[zmichels]

Can you link/share the CTF file?

You mention/reference using code that are not included in your post, so it is difficult to troubleshoot. If you are able to share the CTF file, I think we can at least sort out if something can be done.

[saeadi]

https://files.fm/u/pstgappx69

Thank you very much for your time.

[zmichels]

Seems to be some issue with that link as it does not work for me.

[saeadi]

0pass.zip

How is it now? Is it open?

[kilir]

Hi,
a) with respect to the lines you provide hereyou are smoothing grain(boundaries) but not orientations (so no denoising), but plot orientations again.
b) your data looks like it has quite a lit of pseudosymmetry/misindexing, which probably is associated with a misorientation angle way above the treshold. The EBSD smoothing is grain boundary-aware, hence should not touch those issues.
c) find out what characteristic misorientation is associated with your indexing issue so you mightbe able to get rid of it eventually.
Cheets,
Rüdiger

[saeadi]

Thanks for the advice but I followed the code as shown in the image below. I want to reduce the noise and also make the grains look better.

[kilir]

What you are dealing with is not noise but systematic misindexing.
There's nothing to denoise (or even denoising the data won't be visible/help with ther misindexing).

You may try to find the characteristic misorientation associated with your misindexing and try to recover the data.

Cheers,
Rüdiger

[saeadi]

How can I find misindexing?? Is it possible with mtex? I'm really confused
Thanks for the help

[ralfHielscher]

Hi saeadi,

I took a look at your data and it appears that this ctf file uses a strange definition of Euler angles.
If you do the following

ebsd = EBSD.load('0pass.ctf')
ebsd.rotations = inv(ebsd.rotations)

you get the following result

which looks much more reasonable. It should investigate why how correct the import is this way and why your ctf files behave so strange.
Ralf.

[saeadi]

Yes, I was able to plot this image. But I need a map with clearly defined grains and boundaries.
I only have the initial file in 'ipj' format, which I converted to text format with INCA software, and I want to work with MTEX ​​to continue working, which is how it turned out!