Automatic quantification of spheroidal graphite nodules using computer vision techniques

Abstract

The microstructure of nodular cast irons is characterized by the presence of spheroidal graphite nodules. Metallographic tests may show irregular or degenerated graphite nodules that indicate a reduction in tensile strength of the material as well as its yield limit. This work proposes a computer vision algorithm to estimate the amount of degenerated graphite nodules as well as image analysis necessary to determine the relationship between this quantity of degenerated nodules and the loss of mechanical properties of the nodular cast iron. The proposed algorithm was tested using two cast iron samples, by measuring their microhardness and tensile strength. The results show that the amount of degenerated graphite nodules is inversely proportional to the limit of traction resistance. Sample “A”, of the two samples tested, presented more degenerated nodules than “B” and a lower limit of traction resistance; therefore, it needs less strength to break.