4DSTEM Studio

Electron Microscopy Studio · Reference

EBSD Patterns

Applies to Electron Microscopy Studio 2.1 and later

Patterns browses the patterns of an EBSD or TKD scan, forms virtual images from them, and processes them before indexing. This page lists every control of the part, which is in the EBSD workspace.

The part needs an h5ebsd pattern file. The processing follows kikuchipy; the names in brackets below are kikuchipy's, so a run can be redone in Python.

Calibration

The detector values of the file: they are shown, not edited. The projection centre is in the Bruker convention, as kikuchipy stores it.

Virtual images

Control What it does
Grid 1 is the whole detector, the virtual BSE image. A larger number cuts the detector into that many tiles each way.
Map The image the scan map draws: the mean, a tile, or the colour mix
Red, Green, Blue The tile summed into each channel of the colour mix
Tiles Shows the image of every tile in a pane under the map. A click makes it the map.

In the colour mix each channel is normalised to its median and spread, then the three are rescaled together, as in kikuchipy.

Processing

Each stage is a section with a switch. The pattern on screen shows the stages as they are set now; Process runs them over the whole scan. The stages that need the neighbouring patterns show only once Process has run.

Stage Controls What it does
Static background Remove, Source, Operation, Scale Removes the shading of the detector, which is the same in every pattern (remove_static_background). When the file carries no static background, the mean pattern of the scan stands in.
Dynamic background Remove, Filter, Operation, Std, Truncate Removes the low-frequency intensity of each pattern, a Gaussian blur of it (remove_dynamic_background). An empty Std is kikuchipy's default, the pattern width divided by 8.
Average neighbours Average, Window Averages each pattern with its neighbours in the scan, weighted by the window (average_neighbour_patterns)
NLPAR Average, Radius, λ Averages each pattern with its similar neighbours (pyebsdindex NLPAR). The radius r searches (2 r + 1) × (2 r + 1) patterns. λ is chosen from the noise of the scan.
FFT filter Filter, Pass, cutoffs and widths in pixels Multiplies the spectrum of each pattern by a transfer function (fft_filter). Low-pass smooths the noise.
Intensity Intensity, Percentiles, Std, ÷ √n, Equalise, Kernel, Clip limit Rescales each pattern to the full grey range (rescale_intensity), optionally between two percentiles, or normalises it. Equalise is the contrast-limited adaptive histogram equalisation (adaptive_histogram_equalization).
Size Bin Sums k × k pixels (downsample). The note below says the size of the processed patterns and the binning of the detector that follows.

Pattern, in the Static background section, switches the pattern on screen between the file's own and the processed one.

Indexing uses what you set here

The processing, the background and the geometry belong to Patterns. A change here makes the result in Indexing stale.

Export

Export What it writes
The scan map The map on screen as a PNG, one pixel per scan point
The pattern The pattern on screen as a PNG, one pixel per detector pixel, with its colormap and contrast
The pattern with overlay The pattern with the Kikuchi overlay as drawn: bands, edges, zone axes and labels
The overlay as CSV The band lines (hkl, ends in detector pixels, θ) and the zone axes (uvw, position), for a figure made elsewhere
The header as JSON The scan, pattern and detector header as the reader holds it
The processing Every stage with its parameters in kikuchipy's names, to redo the run in Python

Method

Pattern processing as in kikuchipy 0.13.1; NLPAR as in pyEBSDindex (Brewick et al., Ultramicroscopy 200 (2019) 50). The functions and their tolerances are listed in the parity notes.