Shape analysis of sand particles based on Fourier descriptors

Area of Science:

• Geotechnical Engineering
• Material Science
• Computer Vision

Background:

• Particle shape is crucial for geotechnical material mechanical behavior.
• Quantitative analysis of particle shape, especially for sand, remains challenging.
• Fourier descriptors are established shape descriptors, primarily used in computer vision.

Purpose of the Study:

• To apply Fourier descriptors for comprehensive quantitative analysis of sand particle shape.
• To investigate the statistical relationships between various shape descriptors and Fourier descriptors.
• To explore the utility of Fourier descriptors in discriminating sand samples and reconstructing particle shapes.

Main Methods:

• Collected 600 sand particles from the Hutuo River.
• Calculated shape descriptors: elongation, circularity, convexity, and roughness.
• Utilized Fourier descriptors for shape analysis and statistical correlation.

Main Results:

• Fourier descriptors effectively discriminate sand samples, visualized via Andrews plots.
• A functional relationship was established between Fourier descriptors (Dk) and frequency (k).
• Elongation showed a moderate correlation with the first Fourier descriptor (D1).

Conclusions:

• Fourier descriptors provide a robust method for quantitative particle shape analysis in geotechnics.
• The study demonstrates the potential of Fourier descriptors for sand sample discrimination and shape reconstruction.
• Established correlations offer insights into the relationship between particle morphology and Fourier descriptor parameters.