Digital image correlation in polar coordinate robust to a large rotation

Abstract

Digital image correlation (DIC) is a noncontact optical technique for strain and deformation measurement. In this paper, we propose a novel DIC method in polar coordinate system. Compared with the conventional DIC, the proposed method is robust to a large rotation. In the proposed method, we transfer the rotation in Cartesian coordinate to the translation in polar coordinate. This translation can then be removed by resetting the 0° direction in the polar coordinate. In the proposed method, the shape of both reference and deformed subsets is defined as a circle and these subsets are then described in polar coordinate. The gradient orientation at the seed point or searching point is defined as a new 0° direction in polar coordinate. In order to obtain sub-pixel accuracy, we employ the efficient inverse compositional Gauss–Newton (IC-GN) iteration algorithm. The initial value for iteration obtains the parameters about rotation by considering the gradient orientation error between the gradient orientation at the seed point and that at the search point. Hence, the initial value for iteration is more accurate and the iteration converges fast in the case of a large rotation. Both simulation and experimental work are carried out to verify the validity of the proposed method.