Some Practical Considerations in Finite Element-Based Digital Image Correlation

Bo Wang,Bing Pan,Gilles Lubineau
DOI: https://doi.org/10.1016/j.optlaseng.2015.03.010
IF: 5.666
2015-01-01
Optics and Lasers in Engineering
Abstract:As an alternative to subset-based digital image correlation (DIC), finite element-based (FE-based) DIC method has gained increasing attention in the experimental mechanics community. However, the literature survey reveals that some important issues have not been well addressed in the published literatures. This work therefore aims to point out a few important considerations in the practical algorithm implementation of the FE-based DIC method, along with simple but effective solutions that can effectively tackle these issues. First, to better accommodate the possible intensity variations of the deformed images practically occurred in real experiments, a robust zero-mean normalized sum of squared difference criterion, instead of the commonly used sum of squared difference criterion, is introduced to quantify the similarity between reference and deformed elements in FE-based DIC. Second, to reduce the bias error induced by image noise and imperfect intensity interpolation, low-pass filtering of the speckle images with a 5×5 pixels Gaussian filter prior to correlation analysis, is presented. Third, to ensure the iterative calculation of FE-based DIC converges correctly and rapidly, an efficient subset-based DIC method, instead of simple integer-pixel displacement searching, is used to provide accurate initial guess of deformation for each calculation point. Also, the effects of various convergence criteria on the efficiency and accuracy of FE-based DIC are carefully examined, and a proper convergence criterion is recommended. The efficacy of these solutions is verified by numerical and real experiments. The results reveal that the improved FE-based DIC offers evident advantages over existing FE-based DIC method in terms of accuracy and efficiency.
What problem does this paper attempt to address?