作者: John J. Boyle , Robert B. Pless , Stavros Thomopoulos , Guy M. Genin
DOI: 10.1115/1.4045813
关键词: Algorithm 、 Image warping 、 Numerical differentiation 、 Stereopsis 、 Displacement (vector) 、 Computer science 、 Regularization (mathematics) 、 Triangulation (computer vision) 、 Deformation (meteorology) 、 Image registration
摘要: Estimating strain on surfaces of deforming three-dimensional (3D) structures is a critical need in experimental mechanics. Although single-camera techniques excel at estimating deformation surface parallel to the imaging plane, they are prone artifact for 3D motion because cannot distinguish between out-of-plane and in-plane dilatation. Multiview (e.g., stereo) camera systems overcome this via three-step process consisting of: (1) independent registration, (2) triangulation estimate displacements, (3) estimation. However, existing methods errors associated with numerical differentiation when computing fields from displacement unless regularization schemes used. Such can introduce inaccuracy into Inspired by previous work which combined registration estimation single step 2D images stacks, we developed theory simultaneous image triangulation, multiview system. The does not require fields. We present here theoretical foundations derivation two related implementations approach, discuss their strengths weaknesses.