The purpose of this study was to investigate the applicability of the digital image correlation method, which uses image processing to obtain areal strain distributions, for the observation of rock fractures. In the results of the Brazilian tensile test, the strain increase process just prior to crack initiation was captured when the images were taken at 8,000 fps. The tensile testing of a rubber sheet coated with irregular spots confirmed that the accuracy of the analysis was sufficient, indicating that digital image correlation can be applied for the observation of crack propagation in rocks. By simulating the Brazilian tensile test with a manual mechanical jack, the experiment was conducted on gypsum specimens, and high-resolution images of the experiment were taken. The results of the analysis showed that the cracks started at the points of increased strain by comparing the normal strain distribution in the X direction and the shear strain distribution with the actual crack locations.
It is important to evaluate the initiation and propagation of cracks in a rock mass during the cavity excavation. In the case of hard rock, the fracture initiation occurs in a short time, making it difficult to observe the process. However, recent advances in digital imaging devices have made it possible to capture high-speed, high-resolution images of rock fracture experiments. If the strain distribution in an area can be measured by image processing, it will be possible to track the initiation and progress of the fracture.
Strain gauges have usually been used to measure the strain generated in structures. However, the strain value obtained from a strain gauge is the average strain at one point in one direction; and thus, many strain gauges need to be installed for multipoint measuring. This involves complicated wiring to the measurement device, which causes problems in terms of labor and cost. In addition, as cracks develop or crack widths increase, the strain gauges may break off, making measurement impossible. On the other hand, digital image correlation can obtain the strain distribution over a wider range than strain gauges. It is a simple method requiring only the image of the object to be captured. Digital image correlation (DIC) can measure complex cracks and large deformations. DIC uses a camera to capture the speckle pattern (irregular dots) on a sample surface and measures the displacement and strain distribution from the correlation of images before and after deformation.