As a representative non-interferometric optical technique, the digital image correlation (DIC) can provide full-field displacement and strain measurement for the deformed rocks. However, the standard DIC technique has a limitation in measuring the displacements at the discontinuity and cannot be directly used for identifying the crack mechanism. Thus, a new DIC-based method is proposed for automatically tracing the discontinuities and quantitatively identifying the crack mechanism, i.e. mode I, mode II, and mixed-mode I/II. The new method involves three steps, including displacement measurement from the standard DIC technique, displacement field reconstruction at the discontinuity with the modified subset splitting technique, and post-processing for crack identification and displacement jump measurement. The effectiveness and robustness of the modified subset splitting technique and post-processing method have been verified with the synthetic images and theoretical displacement fields of mode I crack and dislocation. Then, the proposed method is utilized to locate cracks and quantitatively identify the crack mechanism of the initiated cracks in red sandstone containing a single flaw under uniaxial compression. The crack development in the flawed red sandstone specimens is analyzed and the crack types are summarized, in which wing cracks are in mode I, while horsetail cracks and anti-wing cracks are identified as mixed-mode I/II crack. It is shown that the new method avoids some ambiguous identification of crack mechanism and present more objective results.

作为一种代表性的非干涉光学技术，数字图像相关性（DIC）可以为变形的岩石提供全场位移和应变测量。然而，标准的DIC技术在测量不连续处的位移方面具有局限性，并且不能直接用于识别裂纹机理。因此，提出了一种新的基于DIC的方法，用于自动跟踪不连续性并定量识别裂纹机理，即模式I，模式II和混合模式I / II。该新方法包括三个步骤，包括标准DIC技术的位移测量，使用改进的子集分裂技术在不连续处重建位移场，以及裂纹识别和位移跳跃测量的后处理。通过合成图像和I型裂纹和位错的理论位移场，验证了改进的子集分裂技术和后处理方法的有效性和鲁棒性。然后，利用该方法对单轴压缩下含单个缺陷的红砂岩中裂纹的位置进行定位，并定量确定其裂纹的开裂机理。分析了有缺陷的红砂岩试件的裂纹发展并总结了裂纹类型，其中翼型裂纹处于I型，而马尾裂纹和反翼型裂纹则被识别为I / II型混合裂纹。结果表明，新方法避免了对裂纹机理的模糊识别，并给出了较为客观的结果。