Uncertainty quantification for 3D digital image correlation displacement measurements using Monte Carlo method

Abstract

Three-dimensional digital image correlation (3D-DIC) is a leading optical measurement technique for measuring full-field shape, displacement, and deformation of solid materials and structures. However, compared with the well-developed and well-accepted measurement technique, uncertainty quantification (UQ) of 3D-DIC measurements is less advanced and less widely practiced. There still lacks a simple and effective method for quantifying the uncertainty of displacements measured by a specific 3D-DIC system for a specific speckle pattern. This is because the 3D-DIC practice involves a complex and long measurement chain and the propagation of uncertainty is highly nonlinear, which makes the uncertainty quantitation tricky. This work proposes a Monte Carlo-based method to quantify the uncertainty of 3D-DIC displacement measurements. The method is based on a theoretical analysis of random error in image matching and theoretical estimation of camera calibration uncertainty. It can estimate the uncertainty in 3D-DIC displacement measurements without additional experimental procedures. Validation experiments reveal that the estimated uncertainty agrees well with the actual situation. Based on the proposed method, the effects of the speckle pattern quality, subset size, and camera calibration quality on 3D digital image correlation displacement measurement uncertainties are also examined. The results reveal that, compared with image matching, camera calibration has a more pronounced influence on the uncertainty of displacement measurements. The proposed method can be integrated with existing 3D-DIC software to quantify the metrological performance of 3D-DIC measurements and therefore better interpret the measurement results.