A Subpixel Displacement Estimation Algorithm for Digital Image Correlation Based on a Nonlinear Intensity Change Model
X.Y. Liu, Q-C. Tan, L. Xiong and G.D. Liu
The digital image correlation (DIC) method is one of optical metrology for full-field displacement and strain estimation based on digital image processing and numerical computation. The principle of the method is to measure the displacement by comparing the two images of the specimen surface before and after deformation. However, the intensity of a point in the two images in practical applications may change due to the environment condition. In this paper, a nonlinear model for intensity change between reference and deformed images is used to improve the gradient-based algorithm of DIC. Then the closed-form solution of the vector of subpixel displacement is derived from minimizing the digital correlation function of the sum of square difference. The accuracy and robustness for intensity variations of the improved gradient-based algorithm are verified through computer-simulated experiments.
Keywords: Digital image correlation (DIC), laser speckle, gradient-based algorithm, nonlinear intensity change model