An Initial Guess Method for Digital Image Correlation Based on Multi-Level Grid Interpolation and Motion Characteristics Analysis

The Digital Image Correlation (DIC) method is a non-contact measurement technique based on digital image processing, used to capture surface deformations of objects. To reduce the computational load for DIC measurement, this paper proposes a multi-level grid interpolation and motion characteristics analysis initial guess method. In a graph, all subsets that need to be calculated are divided into multiple levels of grids, with the highest spacing at level 0 and decreasing spacing at higher levels. The highest-level grid takes the subset interval set normally. Based on the deformation parameters of low-level grid subsets, the initial deformation parameters of higher-level grid subsets can be estimated. In the image sequence of quasi-static material tests, the motion characteristics of the subsets can be analyzed based on its past positions, and the position of the subset in the next image can be estimated. The initial guess method based on motion characteristics analysis can be used to estimate the deformation parameters of the O-level subsets. The proposed method is validated using images from the classic "DIC Challenge". Compared with the feature point matching based initial value estimation method, the algorithm proposed in this paper can effectively improve the accuracy of overall initial value estimation, and improving the overall execution speed of the algorithm.