This thesis describes research into the automated measurement and classification of the size and shape of aggregate particles. The shape properties of aggregate materials that affect their performance in construction applications are identified, and the current industry standards for measuring these properties are reviewed. Existing approaches to automating the classification of aggregate particle size and shape are critically reviewed, compared and discussed. The design, manufacture and operation of a bench-top device that uses a laser triangulation technique to recover threedimensional data from aggregate particles is described, and the development of novel techniques to classify the shape properties of the aggregate particles based on the acquired three-dimensional data is presented. Opportunities to further extend the work are identified and discussed. The outcomes of this research programme include a number of original and novel contributions to the field of image-based particle analysis. A new three-dimensional mathematical morphology algorithm for the classification of particle angularity is described. This operates on range-intensity images acquired through laser triangulation, and effectively simulates the natural wear process through which particles becomes rounded. A novel 'virtual sieving' algorithm is developed that converts particle size in terms of length, width and height, determined through laser triangulation, into the equivalent sieve sizes used in the aggregates industry. A benchtop prototype system was constructed that uses static laser triangulation to capture range data from the upper surfaces of particles passing along a conveyor. All the algorithms described use the range-intensity image representation produced from this range data as the starting point, thus providing a holistic approach to the automated analysis of particle size and shape.

No description available.

- / -