Computational methods for automated center determination in electron diffraction patterns

Area of Science:

• Materials Science
• Crystallography
• Data Analysis

Background:

• Accurate center detection in electron diffraction patterns is essential for reliable crystallographic analysis.
• Automated processing of diffractograms requires robust and efficient center determination techniques.

Purpose of the Study:

• To evaluate and compare various automated center detection algorithms for electron diffraction patterns.
• To provide guidance on selecting appropriate methods for different types of diffraction data (polycrystalline vs. single-crystal).

Main Methods:

• Comparison of maximum intensity detection, phase cross-correlation, autocorrelation-based detection, pseudo-Voigt profile fitting, and Hough-transform-based detection.
• Evaluation based on accuracy, robustness, speed, preprocessing needs, and applicability to diverse materials.
• Implementation and testing within the open-source Python package EDIfF.

Main Results:

• Phase cross-correlation demonstrated high performance for polycrystalline diffractograms with diffraction rings.
• Pseudo-Voigt profile fitting proved most effective for single-crystal diffractograms with discrete diffraction spots.
• All tested algorithms were integrated into the EDIfF package for automated diffractogram processing.

Conclusions:

• The study offers practical recommendations for choosing center detection techniques in automated electron diffraction analysis.
• The findings contribute to improved data quality and reliability in crystallographic studies.
• The EDIfF package provides a flexible and automated solution for diffractogram center detection.