Study on Fractographical Feature Extraction for Distinguishing Fracture Surface Morphology

The fractography is one of the useful technique for finding the cause of breakage in the engineering parts. However, it is not easy to find an appropriate point on the fracture surface, and also the distinction of fracture surface morphology, such as dimple, river-pattern, striation and so on, needs highly skilled technique. To assist unskilled engineers, there are many studies for quantitative analysis of fracture surface morphology by means of Fractal dimension measurement. It is very interesting that fracture surface possesses fractal structure, but it is still difficult to distinguish the fracture surface morphology by using fractal dimensions measurement. On the other hand, Recurrence plots introduced by Eckmann et al. provide a way to visualize the periodic nature of events such as fracture surface’s characteristic feature. Therefore, by applying the image recognition technique to Recurrence plots obtained from fracture surface, it is expected to see the fractographical feature extraction for distinguishing fracture surface morphology.In this study, in order to recognize the Intergranular fracture on fatigue-fractured spring specimen in a salt water environment, the image recognition technique is applied to Recurrence plots obtained from fracture surface. Since we obtained appropriate judgements for intergranular fracture, it is found that Recurrence plots and image recognition technique could be useful for a computer aided fractographical new technique.