Quantitative shape evaluation of graphite particles in ductile iron

Abstract Ductile irons’ mechanical properties are strongly influenced by the microstructure of metal matrix, presence of impurities and presence of graphite particles. These particles are characterized by their morphological characteristics (e.g. shape, dimension, etc.): the more the shape of the graphite particles is near to a sphere (nodule), the better the mechanical properties (e.g. toughness and crack propagation resistance) are. The standard that is commonly used to characterize the morphology of graphite particles is only semi-quantitative. In this work, an image processing computer procedure is proposed for the characterization of the shape of graphite nodules. Pictures of the metallographic sections acquired by means of a light optical microscope are processed in order to obtain a simpler image representation where the nodules are black objects over a white background (image binarization). On this binary picture each nodule can be easily examined to estimate all the shape parameters of interest, lime area, eccentricity, solidity, etc. Suitable statistics of the shape parameters are adopted to classify the real data of ductile iron specimens.