A new procedure for the automatic measurement of the secondary dendrite arm spacing (SDAS) from microscopic images is presented. The individual primary and secondary dendrite arms are identified through suitable segmentation techniques and clustered in such a way that dendritic structures are obtained suitable for SDAS measurement. The algorithms are applied to two different hypoeutectic aluminum cast alloys, and the quality of the measurements obtained is assessed through a comparison to manually measured SDAS values. A good agreement between the automated measurements and the distribution of manual measurements is found for both cast structures considered. In addition, a decrease in computation time is observed which allows for an increase in measurement density that is used to characterize the microstructures.
Keywords: SDAS; aluminium alloys; image processing; microstructure.