In this first of a two part sequence of papers, 3-D microstructures of Si particle reinforced aluminum matrix composites are computationally constructed by assembling digitally acquired micrographs obtained by serial sectioning. The material samples considered vary in volume fraction and in particle size. Furthermore, equivalent microstructures with actual particles replaced by ellipses (in 2-D) or ellipsoids (in 3-D) are computationally simulated for efficiency. The equivalent microstructures are tessellated by a particle surface based algorithm into a mesh of Voronoi cells. Various 3-D characterization functions are developed to identify particle size, shape, orientation and spatial distribution in the actual materials and to compare with 2-D micrographs. Through this analysis, differences between 2- and 3-D characterization are established. Results indicate that it may not be sufficient to use 2-D section information for characterizing detailed microstructural features like particle shapes, orientations and near-neighbor distances. The second part of this sequence of papers will describe the important relationship of these features to damage evolution in these same materials. This sequence of papers is perhaps one of the first on 3-D physical characterization of the phase and damage structure for this class of materials.