Cleaning the Euler angles
353 views
Skip to first unread message
kasra momeni
Mar 29, 2016, 11:55:28 AM3/29/16
to dream3d-users
Hi,
I have been trying to reproduce Euler angles obtained form EBSD, using Dream3D. However, in the output file from DREAM.3D, there are a lot of noise and most of the features are missed. I was wondering if you could help me on how to clean the Euler angles. I have attached the h5ebsd and the dream3d pipeline files. You probably need to make a few changes in some of the filters that read and write the files to make it compatible with your path. I have also attached an image of the Euler angles obtained using the EBSD machine's software.
Thank you in advance for your help.
Best
Kasra
Michael Jackson
Mar 29, 2016, 12:20:44 PM3/29/16
to kasra momeni, dream3d-users
Who
is your EBSD Machine made by? Oxford? EDAX? Bruker? Someone else? Some
of the angle files that are generated from those machines have the Euler
angles stored in degrees and some in radians. You will need to know
this information in order to get DREAM.3D to correctly generate an IPF
color map. DREAM.3D also does not color by "Euler Angle" as that is a
manufacturer specific coloring scheme.
--
Mike Jackson [mike.j...@bluequartz.net]
Mike Jackson [mike.j...@bluequartz.net]
March 29, 2016 at 11:55 AM
--
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 post to this group, send email to dream3...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dream3d-users/1e467727-b88d-4fa8-9294-2629b61d63e5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
kasra momeni
Mar 29, 2016, 2:07:43 PM3/29/16
to dream3d-users
Hi,
Thank you for your prompt response. The original data were in Oxford ctf format. Where I should specify which unit has been used in DREAM.3D and is there a way to reduce the noise in the final processed data. I used the paraview to visualize the xdmf output files.
best
Michael Jackson
Mar 29, 2016, 2:35:04 PM3/29/16
to kasra momeni, dream3d-users
Kasra,
After looking at your pipeline I believe the first missing filter is the "Convert Angles to Degrees or Radians". Oxford stores the Euler angles in Degrees while DREAM.3D expects them in radians. The second issue is the "Threshold Objects" filter. The data contains 2 phases but are numbered from 0-2 where 0 denotes a scan point that was NOT able to be indexed. I would actually threshold off the "Error" cell data and change the "Threshold Objects" parameters to be "Error = 0". In CTF files when the Error data is NOT equal to 0 then the indexing software could NOT find a solution. Next up is the "Segment Features (Misorientation)". Running a the "IPF Colors" and "Misorientation Colors" filters shows a highly textured sample where all the orientations are fairly close. We tried out the Segment Features filter with a tolerance of 1.5 degrees and was able to extract out ~9900 features. We were not sure what you were really after with this data set but these ideas should be able to get you started. I'll attach some images and our short pipeline for your reference.
After looking at your pipeline I believe the first missing filter is the "Convert Angles to Degrees or Radians". Oxford stores the Euler angles in Degrees while DREAM.3D expects them in radians. The second issue is the "Threshold Objects" filter. The data contains 2 phases but are numbered from 0-2 where 0 denotes a scan point that was NOT able to be indexed. I would actually threshold off the "Error" cell data and change the "Threshold Objects" parameters to be "Error = 0". In CTF files when the Error data is NOT equal to 0 then the indexing software could NOT find a solution. Next up is the "Segment Features (Misorientation)". Running a the "IPF Colors" and "Misorientation Colors" filters shows a highly textured sample where all the orientations are fairly close. We tried out the Segment Features filter with a tolerance of 1.5 degrees and was able to extract out ~9900 features. We were not sure what you were really after with this data set but these ideas should be able to get you started. I'll attach some images and our short pipeline for your reference.
--
Michael A.
Jackson
400 S. Pioneer Blvd
Owner, President Springboro, Ohio 45066
BlueQuartz Software, LLC EMail: mike.j...@bluequartz.net
Voice: 937-806-1165 Web: http://www.bluequartz.net
Fax: 937-746-0783
Owner, President Springboro, Ohio 45066
BlueQuartz Software, LLC EMail: mike.j...@bluequartz.net
Voice: 937-806-1165 Web: http://www.bluequartz.net
Fax: 937-746-0783
March 29, 2016 at 2:07 PM
To view this discussion on the web visit https://groups.google.com/d/msgid/dream3d-users/fc51e315-ad70-44ff-a581-5cd1a53c873c%40googlegroups.com.
March 29, 2016 at 12:20 PM
Who is your EBSD Machine made by? Oxford? EDAX? Bruker? Someone else? Some of the angle files that are generated from those machines have the Euler angles stored in degrees and some in radians. You will need to know this information in order to get DREAM.3D to correctly generate an IPF color map. DREAM.3D also does not color by "Euler Angle" as that is a manufacturer specific coloring scheme.
kasra momeni
Mar 31, 2016, 2:33:04 PM3/31/16
to dream3d-users
Thank you very much for your help. I tried to use the pipeline that you attached but my DREAM3D did not recognize the filters used in the pipeline. I think the file was not properly transferred. However, I tried to implement what you have mentioned in the text. I changed the misorientation tolerance in the "Segment Features" and plotted the results with paraview. They do not match the results obtained by plotting Euler angles using the EBSD machine's software, although I can identify some similar features. I have attached my pipeline, the Euler angles plot from DREAM3D, and the one from the EBSD software. I was wondering whether I missed anything.
With best regards,
Kasra
Sean Donegan
Apr 1, 2016, 9:50:37 AM4/1/16
to dream3d-users
Hi Kasra,
What exact steps are you doing to perform any cleanup in the EBSD software? This is vitally important for determining which filters in DREAM.3D to use for cleanup, if you want to try and get the results to match up. I would also recommend coloring the EBSD software output by grain Id, if possible. The Euler color scheme can be quite ambiguous as to what it means, and if you want to compare to DREAM.3D's feature segmentation, it's probably better to color both data sets by grain Id.
-Sean
Michael Jackson
Apr 28, 2016, 5:13:04 PM4/28/16
to dream3d-users
Hello
Kasra,
The developers of DREAM.3D have identified an issue in the current release of DREAM.3D in the "Read H5Ebsd" filter that can lead to incorrect results. The issue in question is when reading a H5EBSD file where the Euler Angles are represented in Degrees AND the user selects to apply the 'Recommended Transformations". The Euler Angles are silently treated as radians for some internal calculations that transform from Euler angles to an Orientation Matrix and back to Euler Angles. We are correcting the issue in the nightly builds of DREAM.3D. For those working with the released version the only way to work around the issue is to UNCHECK the "Use Recommended Transformations" and then insert into the pipeline the following 3 filters:
+ [Convert Angles to Degrees or Radians](changeanglerepresentation.html)
+ [Rotate Euler Reference Frame](rotateeulerrefframe.html)
+ [Rotate Sample Reference Frame](rotatesamplerefframe.html)
You can look at the recommended transformations from the ReadH5Ebsd filter to get the proper input parameters for your dataset. The bug fix will be in the next official release of DREAM.3D and will soon be incorporated into the nightly builds of DREAM.3D.
The developers of DREAM.3D have identified an issue in the current release of DREAM.3D in the "Read H5Ebsd" filter that can lead to incorrect results. The issue in question is when reading a H5EBSD file where the Euler Angles are represented in Degrees AND the user selects to apply the 'Recommended Transformations". The Euler Angles are silently treated as radians for some internal calculations that transform from Euler angles to an Orientation Matrix and back to Euler Angles. We are correcting the issue in the nightly builds of DREAM.3D. For those working with the released version the only way to work around the issue is to UNCHECK the "Use Recommended Transformations" and then insert into the pipeline the following 3 filters:
+ [Convert Angles to Degrees or Radians](changeanglerepresentation.html)
+ [Rotate Euler Reference Frame](rotateeulerrefframe.html)
+ [Rotate Sample Reference Frame](rotatesamplerefframe.html)
You can look at the recommended transformations from the ReadH5Ebsd filter to get the proper input parameters for your dataset. The bug fix will be in the next official release of DREAM.3D and will soon be incorporated into the nightly builds of DREAM.3D.
--
Mike Jackson [mike.j...@bluequartz.net]
Mike Jackson [mike.j...@bluequartz.net]
April 1, 2016 at 9:50 AM
Hi Kasra,What exact steps are you doing to perform any cleanup in the EBSD software? This is vitally important for determining which filters in DREAM.3D to use for cleanup, if you want to try and get the results to match up. I would also recommend coloring the EBSD software output by grain Id, if possible. The Euler color scheme can be quite ambiguous as to what it means, and if you want to compare to DREAM.3D's feature segmentation, it's probably better to color both data sets by grain Id.-Sean
On Tuesday, March 29, 2016 at 11:55:28 AM UTC-4, kasra momeni wrote:
--
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 post to this group, send email to dream3...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dream3d-users/afa35515-2629-423b-9906-c365806ffd05%40googlegroups.com.
March 29, 2016 at 11:55 AM
Hi,I have been trying to reproduce Euler angles obtained form EBSD, using Dream3D. However, in the output file from DREAM.3D, there are a lot of noise and most of the features are missed. I was wondering if you could help me on how to clean the Euler angles. I have attached the h5ebsd and the dream3d pipeline files. You probably need to make a few changes in some of the filters that read and write the files to make it compatible with your path. I have also attached an image of the Euler angles obtained using the EBSD machine's software.Thank you in advance for your help.BestKasra
Reply all
Reply to author
Forward
0 new messages