Digital image correlation using ring template for 3D reconstruction

Abstract

Three-dimensional reconstruction is one of the most important applications of digital image correlation technology. At present, three-dimensional reconstruction of complex surface is completed by matching images caught by binocular vision measurement system. When three-dimensional objects are projected onto a two-dimensional plane image, various image factors are integrated into single gray level information. In terms of binocular vision measurement system, the images are caught by two cameras at the same time. Because of the different illumination and the stationary angle between two cameras, there might be holistic gray value differences between images caught by two cameras. These differences will affect both image matching precision and final result. In this paper, an improved binocular matching algorithm based on ring template and Newton-Raphson iteration methods is applied. To satisfy the partial convergence of Newton-Raphson iteration method, a ring template is chosen to match the integer pixel location of nodal points. After that, the integer coordinates of nodal points are already estimated which are taken as the initial value of Newton-Raphson iteration method. The morphology reconstruction of complicated surface can be implemented successfully by using algorithm mentioned in this paper. The algorithm can not only reduce the error caused by holistic gray value differences, but also improve the accuracy of reconstruction.