A novel digital image correlation based on a reliable subset strategy for discontinuous displacement measurement

Abstract

Digital image correlation (DIC) is a widely employed method for deformation measurement. However, standard DIC is severely limited in discontinuous deformation measurement, because the shape functions used in standard DIC cannot accurately describe the discontinuous deformation of the subsets. In this paper, a reliable subset strategy is proposed to address discontinuities in DIC. In the implementation of this strategy, SIFT algorithm is utilized to match reference image and target images to obtain the reliable subsets. The reliable subset is a square region selected around matched feature points on reference image, which can avoid the discontinuous deformation regions. Then the reliable subset is correlated to obtain the exact deformation using the Gaussian weighted correlation function. Finally, the deformation parameters of a point of interest (POI) can be obtained by fitting the displacement of reliable subset centers for a fitting region around POI. In cases where discontinuities exist in the fitting region, the direct fitting method may yield substantial errors for obtaining POI deformation parameters. To ensure the accuracy of displacement measurements, the reliable subsets centers in the fitting region are classified into two categories along the crack face, and the most appropriate category is selected for the fitting process. The performance of the proposed method was verified by numerical simulation images and real images. Experimental results show that the proposed method is highly effective in handling displacement discontinuities.