Abstract
A new method to increase the signal-to-noise ratio S/N of electron backscatter patterns (EBSPs) based upon principal component analysis (PCA) is presented. The PCA denoising method is applied to ten scans of EBSPs from the same region of interest of a recrystallised nickel sample acquired with a decreasing S/N, achieved by reducing the exposure time while increasing the camera gain accordingly. That PCA denoising increases S/N in EBSPs is demonstrated by comparing indexing success rates after both Hough and dictionary indexing (HI and DI) of the Ni patterns having undergone one of four processing routes: i) standard static and dynamic background corrections (standard corrections), ii) standard corrections and pattern averaging with the four closest neighbours, iii) standard corrections and PCA denoising, and iv) standard corrections and pattern averaging followed by PCA denoising. Both pattern averaging and PCA denoising increases the indexing success rates for both indexing approaches for the studied Ni scans, with the former processing route providing the better success rates. The best success rates are obtained after pattern averaging followed by PCA denoising. The potential of PCA denoising to reveal additional pattern details compared to standard corrections and pattern averaging is demonstrated in a pattern from an orthoclase (KAlSi3O8) grain in a geological sample. Software code, and the Ni data sets, are released alongside this article as part of KikuchiPy, an open-source software package dedicated to processing and analysis of EBSPs.
Highlights
Electron backscatter diffraction (EBSD) in the scanning electron microscope (SEM) is a technique to characterise crystallographic features [1]
That PCA denoising increases S/N in electron backscatter patterns (EBSPs) is demonstrated by comparing indexing success rates after both Hough and dictionary indexing (HI and DI) of the Ni patterns having undergone one of four processing routes: i) standard static and dynamic background corrections, ii) standard corrections and pattern averaging with the four closest neighbours, iii) standard corrections and PCA denoising, and iv) standard corrections and pattern averaging followed by PCA denoising
There are minor differences in the indexing success rate using a reference kernel (ISRRK) between the two indexing approaches until scan 5 where it drops to 65 % for Hough-transform based indexing (HI), PCA denoising raises it to 94 %
Summary
Electron backscatter diffraction (EBSD) in the scanning electron microscope (SEM) is a technique to characterise crystallographic features [1]. Data collection and indexing is reliable and automated through commercial systems. An illustrative example is contributions to the indexing of electron backscatter patterns (EBSPs), such as dictionary indexing (DI) [2] available in the EMSOFT package, and the ASTROEBSD package [3]. While the latter package and commercial systems in general rely on extraction of Kikuchi bands in patterns using Hough-transform based indexing (HI), the DI approach compares dynamically simulated patterns to the full experimental patterns.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: IOP Conference Series: Materials Science and Engineering
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
