Camera motion-induced systematic errors in stereo-DIC and speckle-based compensation method

Stereo-digital image correlation (stereo-DIC) is now a standard technique for determination of the mechanical properties of materials and structures. In stereo-DIC, cameras are assumed to be motionless after camera calibration, so the three-dimensional (3D) reconstruction can be implemented using pre-calibrated parameters. However, this assumption is not true in some situations, such as drop test, seismic shaking table test and non-laboratory environment. Due to the presence of ground shaking or wind blowing, it's almost impossible to avoid camera motion in these experiments even if mechanical fixing is adopted. Camera motion during the experiment can undoubtedly introduce significant errors on measured results. Generally, camera motion can be divided into pair motion and relative motion. The influence of these two kinds of motion on stereo-DIC measurement is different and worth of separate study. Keep this in mind, in this paper, the influence of camera motion on stereo-DIC is investigated and the effects of pair motion and relative motion are examined separately using precisely controllable simulated experiments. Specifically, by integrating into the stereo matching, the speckle-based compensation method for relative motion is proposed and the accuracy of the compensation methods is analyzed in detail. The reduction of camera motion-induced systematic errors will be helpful for the further applications of stereo-DIC in non-laboratory environments and engineering fields.

Automated summary

Stereo-digital image correlation (stereo-DIC) has become a standard technique for determining mechanical properties of materials and structures. However, in certain situations like drop tests or seismic shaking, camera motion can introduce significant errors on measured results. This study examines the influence of camera motion on stereo-DIC, specifically focusing on pair motion and relative motion, introducing a speckle-based compensation method for the latter to reduce systematic errors.