Research on the Reliability Comparison of DIC and PIV Methods for Soil Body Image Deformation Analysis Based on Preset Sequence Deformation

Image deformation analysis technology is an important means of scientific research and physical experiments in fields such as water conservancy and seismology. DIC (Digital Image Correlation) and PIV (Particle Image Velocimetry) are the most widely used methods. However, existing research lacks a systematic comparison of the performance of the two methods under a unified benchmark, which leads to insufficient basis for selecting their applicability in analyzing complex deformation of rock and soil. Therefore, this paper establishes three types of typical static deformation images rotation, compression and shear, and three types of dynamic sequence images one-dimensional translation, translational shear and dynamic large deformation as preset evaluation criteria. The displacement analysis accuracy, identifiable particle deformation range, and analysis stability of DIC and PIV are compared and analyzed. The results show that: (1) In terms of analysis accuracy and deformation range, DIC benefits from its high-order shape function. Its error ranges in rotation and shear deformation (4%-15%, 4%-40%) are significantly smaller than those of PIV (8%-60%, 11%-63%), indicating that DIC has a stronger ability to handle non-uniform deformation. (2) In terms of analysis stability, the full calculation mode adopted by DIC effectively avoids error accumulation, and its dynamic time history curve is more consistent with the reference datum line. The incremental calculation mode of PIV leads to drift of the results. (3) In strain field inversion, DIC directly solves strain through the shape function, which is superior to the numerical differentiation post-processing used in PIV. In a model test of a sandy slope, DIC captured strain localization phenomena more accurately, with a data loss rate 4.9% lower than that of PIV.