A novel method for measuring discontinuous deformation in digital image correlation based on partition and dividing strategy

• Published in: Engineering fracture mechanics Vol. 204; pp. 185 - 197
• Main Authors: Han, Jing, Pan, Bing
• Format: Journal Article
• Language: English
• Published: New York Elsevier Ltd 01-12-2018; Elsevier BV
• Subjects: Computer simulation; Correlation analysis; Correlation coefficients; Crack propagation; Deformation; Digital image correlation; Digital imaging; Discontinuity; Discontinuous displacement fields; Division; Fracture mechanics; Iterative methods; Mathematical models; Newton-Raphson method; Numerical analysis; Partitions; Shape functions; Fracture mechanics; Digital image correlation; Discontinuous displacement fields; Crack propagation
• ISSN: 0013-7944; 1873-7315
• DOI: 10.1016/j.engfracmech.2018.09.036

•A simple but effective method based on the partition and division strategy is presented to deal with discontinuity issues during DIC analysis.•The method recognizes the discontinuity line using logistic regression and enables line parameters to come out in a few iterations.•A differentiation strategy is adopted according to the relative sizes of the two sub-blocks, deformation parameters of which are retrieved using Newton-Raphson method.•The efficacy of the presented approach is validated using both numerically simulated images and real experimental images.
Digital image correlation (DIC) is the most widely used photomechanics technique for surface deformation measurement. However, standard subset-based DIC restricts its application to continuous deformation, as the regularly-adopted shape functions fail to model discontinuous deformation with displacement jump when discontinuity presents. In this work, a simple but effective method based on the partition and division of interrogated subsets is presented to deal with discontinuity issues during DIC analysis. During the implementation of this method, correlation coefficient of each pixel within interrogated subsets is first calculated between the reference and deformed images to obtain a matrix of correlation coefficients. Then the correlation coefficient matrix is partitioned into two parts through binarization. Subsequently, a dividing line is fitted between the two sub-blocks, which can be considered as the initial discontinuity profile. After that, a differentiation strategy is adopted according to the relative sizes of the two sub-blocks, deformation parameters of which are retrieved using iterative Newton-Raphson method. The proposed method is verified using both numerically simulated images and real experimental images. Experimental results validate the efficacy and accuracy of the presented approach over existing methods in dealing with discontinuity issues.