M.S. in Engineering
Electrical & Computer Engineering
Henry M. Rowan College of Engineering
National Science Foundation
Optical tomography; Aggregates (Building materials)
Electrical and Computer Engineering
The characterization of 3-D shapes of particles in geomaterial aggregate mixtures is important for understanding the micro-mechanics of granular materials. Also, numerical synthesis of 3-D particle shapes from their corresponding shape descriptors is required for developing discrete element models (DEMs) that can be used to predict particle contact-forces, and ultimately the shear strength of the aggregate mixture.
Previous work has shown that Fourier-based 3-D shape descriptors can be constructed for aggregate mixtures, using a statistical combination of 2-D projections. Furthermore, optical tomography methods using the Algebraic Reconstruction Technique (ART) algorithm has proved capable of synthesizing 3-D shapes from 2-D projections, with accuracy comparable to that obtained by an X-ray microtomograph (the "gold" standard).
This thesis extends and revalidates prior work using images obtained from a larger set of geomaterial mixtures - an extensive sand database has been constructed. Inexpensive optical microscopy methods for synthesizing composite 3-D shapes representative of the entire mixture using multiple 2-D images of particles scattered on an image plane, is explored. An optimization technique based on the Euclidean distance metric has been developed for selecting a subset of such 2-D images for synthesis using the ART algorithm. Results demonstrating the success of this technique are shown to depend on the statistics of the particle mixtures. The algorithm is successful in synthesizing particles similar in shape to the optical and X-ray tomography methods for those aggregate mixtures with fairly homogeneous shapes.
Giordano, Patrick Anthony Jr., "Optimization of optical computed tomography techniques for the synthesis of particle aggregate models" (2007). Theses and Dissertations. 777.