Depth-guided DIC for 3D shape, deformation, and strain measurement on discontinuous surfaces

Digital image correlation (DIC), as a powerful vision-based technique for deformation and strain measurement, is widely used in various fields. However, when facing the object with discontinuous surfaces, the results given by DIC are biased due to the inaccurate shape function description. To address it, a depth-guided DIC method is proposed to enable accurate three-dimensional (3D) shape, deformation, and strain measurement on discontinuous surfaces. The relationship between image deformation fields and object's 3D shape is derived, revealing that matching errors is directly related to the 3D geometry of the tested object. Then, fringe projection profilometry (FPP) is employed to acquire pixel-wise 3D shape data, effectively capturing depth steepness and abrupt gradient changes. Finally, depth-guided DIC strategy is developed to split and regenerate subset in DIC and iterative computations are performed on the new-generated subset to obtain accurate results. Experimental results demonstrate that the proposed depth-guided DIC method significantly improves the precision of displacement and strain field analysis for complex 3D surfaces, offering valuable applications in structural mechanics and bioinspired engineering.