AutoDisk: Automated Diffraction Processing and Strain Mapping in 4D-STEM
Development in lattice strain mapping using four-dimensional scanning transmission electron microscopy (4D-STEM) method now offers improved precision and feasibility. However, automatic and accurate diffraction analysis is still challenging due to noise and the complexity of intensity in diffraction patterns. In this work, we demonstrate an approach, employing the blob detection on cross-correlated diffraction patterns followed by lattice fitting algorithm, to automate the processing of four-dimensional data, including identifying and locating disks, and extracting local lattice parameters without prior knowledge about the material. The approach is both tested using simulated diffraction patterns and applied on experimental data acquired from a Pd@Pt core-shell nanoparticle. Our method shows robustness against various sample thicknesses and high noise, capability to handle complex patterns, and picometer-scale accuracy in strain measurement, making it a promising tool for high-throughput 4D-STEM data processing.