Algorithm for fast evaluation of in-plane fiber orientation in reinforced plastics using light microscopy images

Abstract

Methods for evaluating the orientation of carbon fibers in reinforced plastics vary in complexity and are application specific. Here, we report an algorithm that quickly evaluates in-plane fiber orientation based on determining the correlation coefficient of adjacent regions in microscopy images. The result is not the fiber orientation of individual fibers, but the principal fiber orientation of small image areas. This method is applicable to large areas due to its low computation time and captures varying fiber orientations, making it suitable for the study of injection molded samples with complex geometries. A great advantage is that no information about the fibers or the matrix, or their combination, is needed prior to the analysis. This approach is also suitable for samples with surface defects. Demonstrations of this technique are given for Polyamide 6 and Polypropylene with 30 weight % carbon fibers of different fiber lengths.