Automatic speckle region selection for digital image correlation

With the increase in digital image correlation (DIC) applications, it is an important task to realize the fully automatic measurement and provide an automatic deformation measurement sensor. Keeping this in mind, we propose an automatic speckle region selection method for DIC. This method can be divided into three main steps: initial contour selection, active contour computation, and morphological operation. The sum of modulus of the local intensity gradient vector of the local image is used to determine the initial contour, the active contour computation is driven by improved automatic local Gaussian distribution fitting energy, and the morphological closing operation is finally conducted to achieve better robustness. Experimental results show that the root mean squared error of the localization is less than 0.8260 pixels. We further integrate the proposed method into an open-source DIC software, and the fully automatic deformation analysis is realized. The fully automatic deformation measurement will further widen the application scope of DIC, especially in the field of smart robotics and smart city.