This study aims to determine the number of graphite nodules in ductile iron sample according to NBR 6913 and using techniques process and analysis in digital images. This count determines the mechanical properties of the nodular cast iron, which are used for project designs in the field. Therefore, an operator with the aid of an optical microscope does this count traditionally visually. Aiming to reduce the time of determining the amount of nodes and eliminate the maximum possible error in counting, a software was developed using the C language and with the help of the OpenCV library. This library is open source and incorporates optimized Computer Vision techniques. Two approaches were used to count the nodes, one by Region Growing and another using Watershed. Tests were performed and compared with the results obtained by six experts using traditional visual technique to evaluate the proposed methods. The results show that the methods can be incorporated into a system to calculate the number of nodules in nodular cast iron. At the end of this work, it is possible to confirm that the developed software is useful and reliable to nodule count. Therefore is demonstrated its importance to aid this knowledge area.
Digital Image Processing; OpenCV; cast iron; calculation of graphite nodules
