Automatic quantification of filler dispersion in polymer composites

A facile and objective method is introduced to automatically quantify the filler dispersion in polymer composites through image analysis. This method consists of automatic identification of the fillers in the image and a rigorous measurement of the filler dispersion within the space of functions. Compared with previous methods, this method has the advantages of 1) automatically recognizing the fillers, 2) minimizing the subjectivity induced by the inhomogeneity and noise of the background in the images, 3) a mathematically more rigorous definition of the deviation of the filler dispersion from uniformity, 4) a single performance metric reflecting both the distribution and the size of fillers. Both synthetic and real images from model compounds are used to demonstrate the sensitivity of the proposed method to the dispersion and aggregation of fillers. The computed dispersion index shows good agreement with visual observation of synthetic images and mechanical properties of the model compounds. Crown Copyright (C) 2012 Published by Elsevier Ltd. All rights reserved.