Compressive electron backscatter diffraction imaging.

Area of Science:

• Materials Science
• Crystallography
• Computational Methods

Background:

• Electron backscatter diffraction (EBSD) is a key crystallographic characterization technique.
• Current limitations include complex sample preparation, slow acquisition, and beam damage to sensitive materials.
• These factors restrict the quantity and quality of obtainable data.

Purpose of the Study:

• To develop a method for reconstructing incomplete EBSD datasets.
• To address limitations of slow acquisition and beam sensitivity in EBSD.
• To improve the efficiency and applicability of EBSD analysis.

Main Methods:

• Proposed a novel inpainting method using dictionary-learning based beta-process factor analysis (BPFA).
• Subsampled probe positions in EBSD data and reconstructed missing data points.
• Simulated subsampled and noisy EBSD datasets using Gaussian and Poisson noise models.

Main Results:

• Achieved high-quality reconstruction of band contrast and inverse pole figure maps from as little as 10% of probe positions.
• Demonstrated that inpainting un-indexed pixels (zero solution pixel detection) enhances reconstruction quality.
• Showed potential for reconstruction using only 5% of probe positions for inverse pole figure maps.

Conclusions:

• The proposed inpainting method significantly accelerates EBSD analysis.
• This technique extends the application of EBSD to beam-sensitive materials.
• The method offers a pathway to obtain high-quality crystallographic data more efficiently.