Improved DIC algorithm for images affected by motion blur

Digital Image Correlation (DIC) is widely used to measure displacement and strain fields. However, motion blur poses challenges in dynamic applications by increasing measurement uncertainty. This work proposes a novel approach for mitigating motion blur effects directly within the DIC algorithm. Blur estimation is done at the subset level, allowing for handling variable blur amounts throughout the analysed surface. Moreover, the estimation leverages the physical relationship between blur and displacement, with displacement measured through DIC. This leads to an iterative process where blur and displacement are simultaneously refined at each step. In this process, instead of deconvolving blur from the image, the estimated blur is applied to the reference image, preventing deblurring artefacts. Tests have been performed considering 2D and 3D DIC, with uniform and non-uniform blur distributions. In both cases, the proposed method significantly reduces uncertainty in displacement fields and nearly eliminates DIC non-convergence issues for larger blur amounts.