Novel Boundary Edge Detection for Accurate 3D Surface Profilometry Using Digital Image Correlation

Abstract

Digital image correlation (DIC) has emerged as a popular full-field surface profiling technique for analyzing both in-plane and out-of-plane dynamic structures. However, conventional DIC-based surface 3D profilometry often yields erroneous contours along surface edges. Boundary edge detection remains one of the key issues in DIC because a discontinuous surface edge cannot be detected due to optical diffraction and height ambiguity. To resolve the ambiguity of edge measurement in optical surface profilometry, this study develops a novel edge detection approach that incorporates a new algorithm using both the boundary subset and corner subset for accurate edge reconstruction. A pre-calibrated gauge block and a circle target were reconstructed to prove the feasibility of the proposed approach. Experiments on industrial objects with various surface reflective characteristics were also conducted. The results showed that the developed method achieved a 15-fold improvement in detection accuracy, with measurement error controlled within 1%.

Keywords

• digital image correlation (DIC)
• edge detection
• random speckle images
• surface profilometry
• automated optical inspection (AOI)