Recovering the polyhedral geometry of fragments.

Area of Science:

• Geophysics
• Computational Geometry
• Materials Science

Background:

• Rock fragment geometry can be approximated by ideal convex polyhedra.
• The number of faces and vertices of these polyhedra contain crucial geophysical information about fragmentation.
• Current methods for identifying these features rely on subjective visual inspection.

Purpose of the Study:

• To develop a reliable algorithm for identifying the number of faces and vertices of ideal convex polyhedra approximating rock fragments.
• To automate the analysis of rock fragment geometry, moving beyond visual inspection.
• To extract geophysical information from rock fragment shapes.

Main Methods:

• The algorithm processes 3D scans of rock fragments (triangulated polyhedra with numerous faces).
• It employs Gaussian smoothing on spherical histograms to determine key face orientations.
• It systematically simplifies the geometry by identifying key planes, reconstructing the ideal polygon, and removing small faces.

Main Results:

• A novel algorithm capable of reliably identifying the ideal polyhedral approximation of rock fragments was developed.
• Two versions of the algorithm were presented and benchmarked against human measurements on 132 fragments.
• The method successfully identified the ideal polyhedral approximation and enabled tracing the shape evolution of rounded pebbles.

Conclusions:

• The developed algorithm provides a systematic and reliable method for analyzing rock fragment geometry.
• This approach allows for the extraction of valuable geophysical information related to fragmentation processes.
• The method has potential applications in understanding geological processes and material evolution.