8
M. P. Echlin et al.
Specify Microstructural
Information Needed
Select
Detectors /
Imaging Modes
Write
TriBeam
Scripts
Perform
TriBeam
Experiment
Raw Data
Algorithmic
Alignment
Positioner
Alignment
Formulation / Setup
Experiment
Data
Collection
Reconstruction
Analysis
Neighbor
Algorithms
Image
Processing
Thresholds
. . .
Finalized
Dataset
Spatial /
Lens
Distortions
Modality
Registration
Attribute
Thresholds
Erosion /
Dilatation
Biased
Features
Statistical
Classification
Thresholds
Region
Growing
Alignment
Cleanup
Segmentation
Artifact
Removal
Distortion
Correction
Voxelized Data
Mesh Data
Simulation
Average
Values
Microstructural
Descriptors
Distributions
Surface
Mesh
Microstructural
Descriptors
Volume
Mesh
. . .
Empirical
Models
Elastic
FFT
Crystal
Plasticity
. . .
. . .
Fig. 4 The 3D EBSD data that is produced by the TriBeam requires a number of postprocessing steps, which are described schematically here. Each dataset requires somewhat different
parameters; however the core structure of that processing is relatively constant
A series of 3D EBSD TriBeam datasets were collected at various resolutions
and at targeted features, including a fatigue crack initiation site and from a region
where high-resolution digital image correlation (DIC) strain information had been
collected [15, 71]. The characteristics of these René 88DT datasets are listed in
Table 2, as well as an identifying name.
The workflow for acquiring, reconstructing, and analyzing 3D datasets is shown
in Fig. 4. Briefly, this workflow includes defining data collection parameters that
are closely tied to an understanding of the problem to be solved. These parameters
include the 3D resolution necessary to capture the relevant microstructural features,
which imaging modalities are required, or very specific parameters such as EBSD
dwell time for pattern diffraction quality or potential pseudosymmetry complications [72–74]. Reconstruction of the 3D data happens next in the workflow, where
a finalized dataset will be defined for analysis. Slice alignment, data cleanup, image
segmentation, artifact removal, and distortion correction may be performed during
this step. Data cleanup and artifact removal are always rooted in an understanding
of the material via detailed 2D characterization. For instance, a minimum grain
size filter may be applied if it is well-known that grains of a very small size do
M. P. Echlin et al.
Specify Microstructural
Information Needed
Select
Detectors /
Imaging Modes
Write
TriBeam
Scripts
Perform
TriBeam
Experiment
Raw Data
Algorithmic
Alignment
Positioner
Alignment
Formulation / Setup
Experiment
Data
Collection
Reconstruction
Analysis
Neighbor
Algorithms
Image
Processing
Thresholds
. . .
Finalized
Dataset
Spatial /
Lens
Distortions
Modality
Registration
Attribute
Thresholds
Erosion /
Dilatation
Biased
Features
Statistical
Classification
Thresholds
Region
Growing
Alignment
Cleanup
Segmentation
Artifact
Removal
Distortion
Correction
Voxelized Data
Mesh Data
Simulation
Average
Values
Microstructural
Descriptors
Distributions
Surface
Mesh
Microstructural
Descriptors
Volume
Mesh
. . .
Empirical
Models
Elastic
FFT
Crystal
Plasticity
. . .
. . .
Fig. 4 The 3D EBSD data that is produced by the TriBeam requires a number of postprocessing steps, which are described schematically here. Each dataset requires somewhat different
parameters; however the core structure of that processing is relatively constant
A series of 3D EBSD TriBeam datasets were collected at various resolutions
and at targeted features, including a fatigue crack initiation site and from a region
where high-resolution digital image correlation (DIC) strain information had been
collected [15, 71]. The characteristics of these René 88DT datasets are listed in
Table 2, as well as an identifying name.
The workflow for acquiring, reconstructing, and analyzing 3D datasets is shown
in Fig. 4. Briefly, this workflow includes defining data collection parameters that
are closely tied to an understanding of the problem to be solved. These parameters
include the 3D resolution necessary to capture the relevant microstructural features,
which imaging modalities are required, or very specific parameters such as EBSD
dwell time for pattern diffraction quality or potential pseudosymmetry complications [72–74]. Reconstruction of the 3D data happens next in the workflow, where
a finalized dataset will be defined for analysis. Slice alignment, data cleanup, image
segmentation, artifact removal, and distortion correction may be performed during
this step. Data cleanup and artifact removal are always rooted in an understanding
of the material via detailed 2D characterization. For instance, a minimum grain
size filter may be applied if it is well-known that grains of a very small size do