In this work, the mixing quality of a multi-component blend of particles is determined using Multivariate Image Analysis of RGB images combined with the Grey Level Co-occurrence Matrix (GLCM) texture analysis. The mixing dynamics and flow behavior are studied in terms of mixing homogeneity and time to achieve equilibrium for binary, ternary and quaternary mixtures of equal size particles. The parameters studied are rotational speed, filling ratio, initial powder composition and particle sizes. Using the experimental data in the rolling regime, a second order model is proposed to predict the mixing curve as a function of time. (C) 2010 Elsevier Ltd. All rights reserved.