Neighbour orientation correlation

[Al Garner]

Hi guys,

I have a large 3D EBSD dataset and have some small (<1.5deg) misorientation between slices due to small errors in the stage movements during the acquisition. I am trying to clean these up using the "neighbour orientation correlation" filter, however it does not appear to be working. Ideally, I would like to clean up these small, artificial misorientations but without removing too much of the "real" deformation (say above 2deg misorientation).

I am unsure how the filter works with Oxford data, as we have the MAD in place of the confidence index, which gives us more confidence in the result he lower it is, how does this work with the "minimum" value set in the parameters?

The other potential issue is as these misorientations result from misalignment of the sample, the "confidence index" is still high for these values (or in the case of MAD, low), 

Is it possible to use this filter for Oxford data? I have tried a lot of different values now and don't seem to see any change in the data. I have attached an XY and YZ slice so you can see the issue and here is a link to a small subset of the CTFs in case you want to try it out.

Thanks in advance, 

Al

[Michael Jackson]

Unfortunately that filter was designed with EDAX data in mind and we have not modified the filter to work with Oxford data. You could potentially use the “Array Calculator” filter to generate a *confidence Index* array by somehow rescaling and inverting the MAD values to go between 0 and 1. That might work.

 

--

Michael Jackson | Owner, President

      BlueQuartz Software

[e] mike.j...@bluequartz.net

[w] www.bluequartz.net

--
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/6b3643d0-68a2-4903-8fa1-6c9aa9c6eaden%40googlegroups.com.