Crack recognition and reconstruction with coarse–fine discontinuous digital image correlation

Abstract

The recognition of structural cracks and reconstruction of displacement fields around cracks are challenging tasks in engineering. By improving the standard digital image correlation (DIC) method and applying it in measuring the discontinuity of cracks, this study presents a coarse–fine search strategy based on a discontinuous DIC for rapid crack recognition and accurate reconstruction of the displacement fields. The technique introduces a split-subset model with a vector that represents the direction of crack opening to overcome the limitation of the standard DIC in analyzing discontinuous areas. A specially designed coarse–fine search strategy is applied to support the calculation model for achieving higher accuracy. The proposed approach is validated using numerically synthesized crack images as well as images obtained from notch tensile experiments in the laboratory. The results show that the proposed method can recognize the crack faces; in addition, it works well for reconstructing the displacement fields of both continuous and discontinuous (cracked) areas with average absolute errors of approximately 0.02. The proposed technique is also valid for slightly noised images.