3D-DIC camera self-motion compensation based on automatic selection of reference features

To improve the displacement measurement accuracy of digital image correlation (DIC) affected by camera self-motion, a compensation method incorporating automatic selection of reference features and depth information constraints was proposed. Traditional motion compensation methods suffer from poor performance due to the improper selection of reference feature points when background motion interference exists. In the proposed method, robust separation between stationary and moving interference regions in the scene was first achieved by performing grid segmentation on temporal difference images. Then, feature points were screened within the stationary region, and those with significant motion were eliminated. Finally, by combining the depth information of the feature points, the set used to estimate the camera self-motion model was further optimized, improving the accuracy and robustness of the camera self-motion compensation. The results showed that compared with the camera self-motion compensation method without screening feature points, the proposed method reduced the root-mean-square error (RMSE) of displacement measurement by 73.5% after camera self-motion compensation. The proposed method better suppressed the interference of camera self-motion on DIC measurement accuracy. These findings contribute to the application of 3-D DIC technology in complex environments.