Dataset Open Access
Ten electron backscatter diffraction (EBSD) datasets from a recrystallised, polycrystalline sample of Nickel. The datasets, each comprising 29 800 patterns, were collected from the same region of interest (ROI) with EBSD camera gains varying from 0 dB to 24 dB (camera maximum). A backscattered electron image of the ROI is also included.
The data was acquired in order to study denoising of EBSD patterns, more specifically the increased signal-to-noise ratio obtained after principal component analysis followed by dimensionality reduction, as presented in H W Ånes, J Hjelen, B E Sørensen, A T J van Helvoort, K Marthinsen "Processing and indexing of electron backscatter patterns using open-source software", IOP Conf. Ser.:Mater. Sci. Eng. (2019) (submitted, as part of the proceedings of EMAS 2019 - 16th European Workshop on Modern Developments and Applications in Microbeam Analysis in Trondheim, Norway). However, the data is released with the hope that it can be used to compare the performance of denoising methods in general.
The datasets, Pattern.dat, are stored in the NORDIF (binary) format, with the top-left pixel in the top-left pattern as the first byte, and the bottom-right pixel in the bottom-right pattern as the last byte. They can be opened in for example the open-source Python package KikuchiPy (https://github.com/kikuchipy/kikuchipy). Assuming Python 3 or above and the package is installed, the patterns in scan 1 can be read and plotted with the following commands:
import kikuchipy as kp s = kp.load('/path/to/nickel_scan_gain/scan1_gain0db/Pattern.dat') s.plot()