Position-based intensity change model for illumination variation in digital image correlation

Abstract

The illumination variation exerts a significant influence on the accuracy and efficiency of digital image correlation (DIC) technology. However, the existing intensity change models assume that all the pixels in the reference subset have the same gray-level intensity change parameters, which cannot accurately reflect the effect of illumination variations. We pay attention to the position effect on intensity variations and propose a position-based intensity change model to study the illumination variation in DIC. The proposed model combines the gray-level intensities, the illumination variation, and the position parameters of a reference subset. The forward additive Gauss–Newton algorithm with a first-order shape function is applied to evaluate the feasibility and effectiveness. Simulation and experimental results verify that the proposed model has higher accuracy in different illumination conditions as compared with the existing second-order intensity change model.