grain2d.merge

A method of grain2d.

the grains joined into one along given boundaries, by pairs of ids, by a label per grain, below a misorientation angle, or every inclusion into the grain around it; MATLAB’s merge

Syntax

merged = grains.merge(grains.boundary['Forsterite', 'Forsterite'])

merged = grains.merge(np.array([[1, 2], [5, 9]]))

merged = grains.merge(grains.neighbors(matrix=True))

merged = grains.merge(threshold=5*degree)

merged = grains.merge(inclusions=True, maxSize=5)

merged = grains.merge(inclusions=True, host=grains.numPixel > 100)

parentId = grains.merge(threshold=5*degree, testRun=True)

Description

Two grains go into one when a criterion joins them, directly or through a chain of other grains. Every criterion given applies at once: pairs or labels, a threshold and inclusions together join what any of them joins.

Grains of two phases are merged only by an inclusion, which then takes the phase and the orientation of its host. A pair of two phases is passed over, and a label spanning phases joins the grains of each phase apart, where MATLAB joins them and gives the merged grain the largest phase id.

The pixel counts of the parts add up, the mean orientation is their mean weighted by the pixels (MATLAB’s calcMeanOrientation, always done here) and GOS their weighted mean. The boundary between the parts becomes inner boundary. The merged grains carry the ids of their parts as childId, and the grains passed get parentId, the id of the merged grain each went into, which MATLAB returns as a second output. MATLAB’s merge at a triple point list, grains.merge(tP), is not provided.

Input

Keywords

Output

Example

Join one neighbouring pair of reconstructed Forsterite grains and inspect the parent id mapping. The small sample is downloaded and cached on first use.

import numpy as np
import matplotlib.pyplot as plt
from mtex import mtexdata, calcGrains, degree, plot, hold

ebsd = mtexdata('small', verbose=False)['Forsterite']
grains = calcGrains(ebsd, angle=10*degree, minPixel=5)
pair = grains.neighbors()[:1]
parent = grains.merge(pair, testRun=True)
merged = grains.merge(pair)
print('Merged grain ids:', pair.tolist())
print('Grain count before and after:', grains.size, merged.size)
print('Preview agrees with parentId:', np.array_equal(parent, grains.parentId))
print('Total pixel count preserved:', int(grains.numPixel.sum()) == int(merged.numPixel.sum()))
plot(grains.boundary, lineColor='lightgray', lineWidth=3)
hold(True)
plot(merged.boundary, lineColor='black', lineWidth=1)
hold(False)
plt.show()

The pair is chosen only to demonstrate the operation; an analysis should select boundaries using a physical criterion. Grey traces show the original network and black the merged outlines. parentId is stored on the original grains. Update any pixel-to-parent mapping explicitly before using the merged grains with EBSD measurements; merge does not rewrite ebsd.grainId.

Merged grain ids: [[1, 2]]
Grain count before and after: 15 14
Preview agrees with parentId: True
Total pixel count preserved: True

See also

grain2d.isInclusion, grain2d.neighbors, calcGrains, GrainMerge, parentGrainReconstructor

Used on

LowLevelParentGrainReconstruction, MaParentGrainReconstructionAdvanced

MATLAB’s grain2d.merge

@article{bachmann2010mtex,
  author  = {F. Bachmann and R. Hielscher and H. Schaeben},
  title   = {Texture Analysis with MTEX - Free and Open Source Software Toolbox},
  journal = {Solid State Phenomena},
  volume  = {160},
  pages   = {63-68},
  year    = {2010},
  doi     = {10.4028/www.scientific.net/SSP.160.63},
  url     = {https://doi.org/10.4028/www.scientific.net/SSP.160.63}
}