A QUANTITATIVE IMAGE-ANALYSIS METHOD FOR CHARACTERIZING MIXTURES OF GRANULAR-MATERIALS

Automated image analysis is used to characterize the structure of undi sturbed granular mixtures. These mixtures, produced from small-scale m ixing experiments, are preserved using solidification techniques, slic ed, and scanned using a video camera to produce a digital image. Image analysis is used to determine the mean, mode, standard deviation, var iance, and skewness of-local composition for a large number of locatio ns in the mixtures, providing an effective method for performing a det ailed quantitative characterization of the mixture structure. The digi tal data are then used to simulate random sampling from a powder blend er. Sampling parameters, such as number of samples, sample size, and l ocation of the samples, are varied to assess their effect on the descr iptive statistics. For partially mixed systems such as those examined here, results depend strongly on the number of samples used in the ana lysis, but they are largely independent of sample size. The solidifica tion techniques, image analysis methods, and numerical algorithms for quantifying features of the partially mixed structures are described i n detail.