Selecting grains containing special boundaries

[Dorian Depriester]

Hello there!
  I would like to characterize few grains which display special boundaries (twinning, actually). Those grains have been constructed from EBSD data. First, I want to extract those grains in a new grainSet, and then perform a texture analysis on it.
I manage to get the faces corresponding to twin planes with:

rot19c=rotation('axis',vector3d(Miller(0,1,-1,0)),'angle',86.98*degree);
f=specialBoundary(grains,rot19c);

My idea was to merge the grains considering those boundaries and then try a boolean operation to remove the non-merged grains. But due to the large number of orientations of my map (over 130,000), when I type:

[merged_grains,I_PC] = merge(grains,rot19c);

I get the following error:

 Maximum variable size allowed by the program is exceeded.

Is there any other (simpler) way to extract those special grains? (using there locations, for instance...)

Thank you very much.

[Dorian Depriester]

Finally, I got rid of this problem by manually selecting my twinned grains (fortunately, the few ones...). But instead of using their locations, I have used their indexes, after displaying them:

centers=centroid(grains);
figure
plot(grains)
for i=1:size(centers,1)
    text(centers(i,1),centers(i,2),sprintf('%i',i) )
end

The 'text' command allows us to zoom on the map without changing the font size.

I'm not that proud of this means, but it solved my problem!

Regards.

[Ralf Hielscher]

Hi Dorian,

thank you for this nice question. Your first step to compute the special boundaries was correct. 

rot19c=rotation('axis',vector3d(Miller(0,1,-1,0)),'angle',86.98*degree);

f=specialBoundary(grains,rot19c);

Having the special boundaries you want to know to which grains they belong. This relationship is stored in a huge sparse incidence matrix which can be extracted from the grain variable.

I_FG = get(grains,'I_FG');

This matrix has as many rows a grain edges and as many columns as grains. A one indicates that a edge belongs to a grain. Thus all we have to do is look at the rows indicated by f and find all columns that have a one there. This is done by the command

twin_ids = find(any(I_FG(f,:),1));

Finally, you can extract your grains by

twins = grains(twin_ids)

Hope this helps,

Ralf.

[Dorian Depriester]

Thank you very much for your quick answer, it really helped. This way is much easier (and smarter) than mine !
This is a very convenient feature.

Best regards.