Techniques have been developed for analyzing a series of SEM and X-ray
images of a cementitious material to quantitatively characterize its
microstructure. For both fly ashes and cement powders, decision trees
have been presented for determining the mineralogical phase to assign to
each pixel comprising the two-dimensional image. Processed images are
conveniently analyzed to estimate phase volume fractions, phase surface
area fractions, and phase two-point correlation functions. The quantitative
data provided by these techniques is of paramount importance, both for
materials characterization and for use as input into computer-based
microstructural models of these complex materials.