Analysis of displacement fields with large deformations using an improved spectral digital image correlation method

Abstract

In this paper, the systematic study of the analysis and measurement of displacement fields with large deformations based on an improved spectral digital image correlation (SDIC) method is conducted. As a shape function including displacement gradients has been successfully introduced, the SDIC method can deal with deformation issues involving different strains. We use the large image Fast Fourier Transform (FFT) to calculate full-field gray gradients instead of the subset image FFT. Therefore, all the gray gradients on image grids can be computed before formation of the Hessian matrix and the iterative optimization process. This can improve the efficiency and accuracy of the SDIC method. To analyze deformation with large strain, an easy-to-use and effective initial displacement estimation method for the seed point is introduced based on the correlation searching technique. The final displacements and displacement gradients of the region of interest (ROI) are then calculated by the subsequent iterative optimization algorithm, which can converge when the initial displacement error is within the range of ± 3 pixels. A set of images with noise ranging from 0 to 4 standard deviations and strains ranging from 0 to 0.6 are simulated and analyzed. The error analysis confirmed that the improved SDIC method performs well in anti-noise and can accurately analyze the displacement field. Finally, the silicone rubber specimens were tested in compression, tension and bending and analyzed to further confirm the validity of the improved SDIC method in the measurement of large deformations.