Error 'reshape'?

[Surya-Cal]

Hello community,

I am trying to process an EBSD dataset that is > 10 mm in length. However, during grain reconstruction, the script produces errors and the resulting grains appear artificially extrapolated into hexagonal shapes, which are not realistic. This does not seem to be a purely MTEX-related error, and since I am very new to the MTEX community, I am not sure what type of issue this might be. I would be really happy to if someone could help.

I am using MATLAB2025b and MTEX version 6.0.0.

My script is provided below:

%% crystal symmetry
CS = {... 
  'notIndexed',...
  crystalSymmetry('2/m', [9.908 18.218 5.3014], [90,104.591,90]*degree, 'X||a*', 'Y||b*', 'Z||c', 'mineral', 'Actinolite', 'color', [0.38, 0.80, 0.42]),...
  crystalSymmetry('-32/m', [4.991 4.991 17.062], [90,90,120]*degree, 'X||a', 'Y||b*', 'Z||c', 'mineral', 'Calcite', 'color', 'blue'),...
  crystalSymmetry('-1', [8.178,12.870,14.188],[93.508, 115.908, 90.658]*degree, 'X||a*', 'Z||c','mineral', 'Feldspar', 'color', [0.8500 0.3250 0.0980]),...
  crystalSymmetry('4/mmm',  [4.587 4.587 2.954],[90, 90, 90]*degree, 'X||a*','Y||b','Z||c','mineral', 'Rutile', 'color', 'magenta'),...
  crystalSymmetry('-1', [8.18 12.87 93.51],[93.51, 115.91, 90.66]*degree, 'X||a*', 'Z||c','mineral', 'Bytownite', 'color', [0.8500 0.3250 0.0980]),... % CIF - 9011200
  crystalSymmetry('2/m',  [5.37 9.30 14.25],[90, 96.28, 90]*degree, 'X||a*','Y||b','Z||c','mineral', 'Clinochlore', 'color', [0.4940 0.1840 0.5560]),...
  crystalSymmetry('2/m',  [4.587 4.587 2.954],[90, 90, 90]*degree, 'X||a*','Y||b','Z||c','mineral', 'Epidote', 'color', [0.4940 0.1840 0.5560]),...
  crystalSymmetry('-1', [8.178,12.870,14.188],[93.508, 115.908, 90.658]*degree, 'X||a*', 'Z||c','mineral', 'Albite', 'color', [0.8500 0.3250 0.0980])};

%% Miller Indices
h_act = [Miller(1,0,0,CS{2},'hkl'), Miller(0,1,0,CS{2},'hkl'),Miller(0,0,1,CS{2},'hkl')];
h_cal = [Miller(0,0,0,1,CS{3},'uvtw'), Miller(1,1,-2,0,CS{3},'uvtw'), Miller(0,1,-1,2,CS{3},'uvtw'), Miller(1,0,-1,4,CS{3},'uvtw'),Miller(0,1,-1,8,CS{3},'uvtw'),Miller(1,0,-1,0,CS{3},'uvtw')];
h_fld = [Miller(1,0,0,CS{4}), Miller(0,1,0,CS{4}), Miller(0,0,1,CS{4})];
h_rut = [Miller(1,1,0,CS{5}), Miller(0,1,0,CS{5}), Miller(0,0,1,CS{5})];
h_chl = [Miller(1,0,0,CS{7},'hkl'), Miller(0,1,0,CS{7},'hkl'),Miller(0,0,1,CS{7},'hkl'),Miller(1,0,1,CS{7},'hkl')];
h_epi = [Miller(1,0,0,CS{8}), Miller(0,1,0,CS{8}), Miller(0,0,1,CS{8})];
h_alb = [Miller(1,0,0,CS{9}), Miller(0,1,0,CS{9}), Miller(0,0,1,CS{9})];
h_byt = [Miller(1,0,0,CS{6}), Miller(0,1,0,CS{6}), Miller(0,0,1,CS{6})];

%% plotting convention
setMTEXpref('xAxisDirection','east');
setMTEXpref('zAxisDirection','north');

%% Specify File Names
fname = 'XXX.crc';

%% Import the Data
ebsd = loadEBSD_crc(fname,CS,'interface','crc',...
    'ColumnNames', { 'Phase' 'x' 'y' 'Euler 1' 'Euler 2' 'Euler 3' 'Mad' 'BC'},...
    'convertEuler2SpatialReferenceFrame');
% Rotation 
ebsd = rotate(ebsd,rotation('axis',vector3d.X,'angle',180*degree));

%% Simple grain construction and cleanup (probably want to modify)
ebsd = ebsd.gridify;
[grains,ebsd.grainId,ebsd.mis2mean] = calcGrains(ebsd,'angle',[1 10]*degree,'Boundary','tight');

%% Remove potentially spurious "grains"
grains = grains(grains.grainSize>5);

% reset the EBSD data to filtered grains
ebsd = ebsd(grains);

%% Recompute grain boundaries with the remaining EBSD data
[grains,ebsd.grainId,ebsd.mis2mean] = calcGrains(ebsd,'angle',[1 10]*degree,'Boundary','tight');

%% smooth jagged/stepped grain boundaries
grains = smooth(grains,5,'MoveTripleJunctions');

%% Reconstruction of grains
notIndexed = grains('notIndexed');

% Show 'notIndexed' area
plot(notIndexed,log(notIndexed.grainSize ./ notIndexed.boundarySize))
mtexColorbar

% the "not indexed grains" we want to remove
toRemove = notIndexed(notIndexed.grainSize ./ notIndexed.boundarySize<4);

% now we remove the corresponding EBSD measurements
ebsd(toRemove) = [];

% and perform grain reconstruction with the reduces EBSD data set
[grains,ebsd.grainId,ebsd.mis2mean] = calcGrains(ebsd);

plot(grains,'lineWidth',0.7)

%% Filtering
F = halfQuadraticFilter;
F.threshold = 1.5*degree;
F.eps = 1e-2;
F.alpha = 0.01;
ebsd = smooth(ebsd('indexed'),F,'fill',grains);

Error:

Error using [reshape](matlab:matlab.lang.internal.introspective.errorDocCallback('reshape'))
Number of elements must not change. Use [] as one of the size inputs to automatically calculate the appropriate size
for that dimension.

Error in [subsind](matlab:matlab.lang.internal.introspective.errorDocCallback('subsind', '/Users/bhupeshmeher/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/private/subsind.m', 63)) ([line 63](matlab: opentoline('/Users/bhupeshmeher/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/private/subsind.m',63,0)))
    ind = ind & reshape(sub,size(ind));
                ^^^^^^^^^^^^^^^^^^^^^^
Error in [indexing](matlab:matlab.lang.internal.introspective.errorDocCallback('grain2d/subsref', '/Users/bhupeshmeher/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/subsref.m', 61)) ([line 61](matlab: opentoline('/Users/bhupeshmeher/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/subsref.m',61,0)))
    ind = subsind(grains,s(1).subs);
          ^^^^^^^^^^^^^^^^^^^^^^^^^
Error in [L1_Map_01_LoadMTEX](matlab:matlab.lang.internal.introspective.errorDocCallback('L1_Map_01_LoadMTEX', '/Users/bhupeshmeher/PhD Project/PROJECT_V_Actinolite/5_EBSD/Location_01/EBSD_Bochum/Map01_Hinge/L1_Map_01_LoadMTEX.m', 97)) ([line 97](matlab: opentoline('/Users/bhupeshmeher/PhD Project/PROJECT_V_Actinolite/5_EBSD/Location_01/EBSD_Bochum/Map01_Hinge/L1_Map_01_LoadMTEX.m',97,0)))
toRemove = notIndexed(notIndexed.grainSize ./ notIndexed.boundarySize<4);
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Output map (please look at the grain shapes):

Thank you very much in advance!
Surya

[kilir]

Hi
what do you plan to achieve with this line

% the "not indexed grains" we want to remove
toRemove = notIndexed(notIndexed.grainSize ./ notIndexed.boundarySize<4);

?
Cheers,
Rüdiger

[Surya-Cal]

Hi
I comment on the following two lines.

% the "not indexed grains" we want to remove
%toRemove = notIndexed(notIndexed.grainSize ./ notIndexed.boundarySize<4);
% now we remove the corresponding EBSD measurements
%ebsd(toRemove) = [];

However, the error persists. Is it because of the larger file size? How can I handle such a large dataset?
My memory is 24 GB.

Error using [false](matlab:matlab.lang.internal.introspective.errorDocCallback('false'))
Requested 2597177x80643 (195.1GB) array exceeds maximum array size preference (24.0GB) and might cause MATLAB to become
unresponsive.

Thank you very much in advance,
Surya

[kilir]

Ok, now, at which line exactly do you get that error?

[Surya-Cal]

Hi,
The following are the errors:

Error using [false](matlab:matlab.lang.internal.introspective.errorDocCallback('false'))
Requested 2597177x80643 (195.1GB) array exceeds maximum array size preference (24.0GB) and might cause MATLAB to
become unresponsive.

Error in [grain2d/checkInside](matlab:matlab.lang.internal.introspective.errorDocCallback('grain2d/checkInside', '/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/checkInside.m', 72)) ([line 72](matlab: opentoline('/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/checkInside.m',72,0)))
isInside = false(size(xy,1),length(grains));
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Error in [grain2d/checkInside](matlab:matlab.lang.internal.introspective.errorDocCallback('grain2d/checkInside', '/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/checkInside.m', 61)) ([line 61](matlab: opentoline('/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@grain2d/checkInside.m',61,0)))
  isInside = grains.checkInside(xy+repmat(uc(1,:),size(xy,1),1),'includeBoundary');
             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Error in [EBSD/fill](matlab:matlab.lang.internal.introspective.errorDocCallback('EBSD/fill', '/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@EBSD/fill.m', 55)) ([line 55](matlab: opentoline('/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@EBSD/fill.m',55,0)))
isInside = checkInside(grains,ebsd.subSet(nanId));
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
Error in [EBSD/smooth](matlab:matlab.lang.internal.introspective.errorDocCallback('EBSD/smooth', '/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@EBSD/smooth.m', 57)) ([line 57](matlab: opentoline('/Users/suryasahoo/Documents/Software/mtex-6.0.0/EBSDAnalysis/@EBSD/smooth.m',57,0)))
if check_option(varargin,'fill') || check_option(varargin,'extrapolate'), ebsd = fill(ebsd,varargin{:}); end
                                                                                 ^^^^^^^^^^^^^^^^^^^^^^
Error in [L1_Map_01_LoadMTEX](matlab:matlab.lang.internal.introspective.errorDocCallback('L1_Map_01_LoadMTEX', '/Users/suryasahoo/PhD Project/PROJECT_V_Actinolite/5_EBSD/Location_01/EBSD_Bochum/Map01_Hinge/L1_Map_01_LoadMTEX.m', 114)) ([line 114](matlab: opentoline('/Users/suryasahoo/PhD Project/PROJECT_V_Actinolite/5_EBSD/Location_01/EBSD_Bochum/Map01_Hinge/L1_Map_01_LoadMTEX.m',114,0)))
ebsd = smooth(ebsd('indexed'),F,'fill',grains);
       ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Thank you in advance
Surya

[kilir]

Hi,
the number 2597177x80643 arises from number_of_pixels_to_be_filled x number_of_grains. So in short and currently expensive terms, yes, you do not have a enough memory.

Various ways to overcome this.
a) you can try to reconstruct your map piece-wise and recombine the data afterwards. However, unrealistic grains remain unrelatic in your common procedure.

b) You have quite poor indexing and many 1 (or only few) pixel grains and you already observed that obviously constructing grains over those poorly indexed areas results in unrealistic grain shapes. So do not construct grains across those non-indexed areas (keep 'notIndexed'' in your dataset). That would imply that your grain shapes are still unrealistic but this time not due to reconstruction but due to measurements.

c) middle ground: eliminate ebsd points in your dataset forming 1 or 2,.. px grains (as large as you feel that they are not trustworthy) and compute grains using a gridified ebsd dataset.

Example:

mtexdata forsterite

% compute grains without voronoi tesselation
[grains, ebsd.grainId] = ebsd.calcGrains('unitCell')

% delete data of grains which are too small to be meaningful
ebsd = ebsd(grains(grains.numPixel>5))

%% inspect
% white areas may be incoporated into grains
plot(ebsd)
hold on
plot(ebsd('notIndexed'),'FaceColor','k')
hold off

%% reconstruct using alphaShapes
ebsd = ebsd.gridify
[grains, ebsd.grainId] = ebsd.calcGrains

%% inspect again
plot(ebsd)
hold on
plot(ebsd('notIndexed'),'FaceColor','k')
plot(grains.boundary,'lineColor','r')
hold off

Not that the 'minPixel' option on EBSDsquare is currently not supported, otherwise you could have done this in one run - hence the first step of grain reconstruction.

d) Acquire the data / use material files such that you have better indexing.

Hope this helps.
Cheers,
Rüdiger