Mesh-Based Shape Measurements with Stereocorrelation

Abstract

A mesh-based framework is developed by extending global stereocorrelation techniques to faceted surfaces with three-noded elements. A two-step self-calibration procedure is followed to determine the projection matrices of the stereo-rig and to update the nominal surface model to match the surface of interest. To prove the feasibility of mesh-based stereocorrelation, two different test parts are analyzed with the present techniques and compared to already validated optical procedures. Keywords Calibration • DIC • finite element discretization • photogrammetry • stereocorrelation 1 Introduction Geometries of manufactured parts are designed to fulfil functional requirements such as assembly constraints and fatigue life. It is therefore essential to directly control the manufacturing process.

On the analysis of canvas wrinkling via isogeometric stereocorrelation

“…In global stereocorrelation, there are two ways of dening the surface of interest. First low order shape functions may be considered (e.g., meshes made 70 of triangular (Dubreuil et al, 2016) or quadrilateral elements). Second higher order interpolations may be considered (e.g., Bézier or NURBS patches Dufour et al, 2015a)).…”

“…The aim of the present work is to use global approaches to stereocorrelation Dufour et al, 2015a;Dubreuil et al, 2016), which consist of measuring 3D shapes and surface displacements when registrations are performed over the whole surface of interest and not locally as in the afore-30 mentioned approaches (Khennouf et al, 2010;Dureisseix et al, 2011;Malowany et al, 2014). Consequently, if some contrasted zones exist, the shape and displacement interpolations will allow the sought information to be assessed everywhere.…”

A modal approach for shape defect measurement based on global stereocorrelation

“…However, the output is a point cloud, which involves a complex and time-consuming post-processing step. A second approach, similar to global approaches adopted in 2D-DIC (2D Digital Image Correlation) [20,21] consists in mathematically modeling the studied surface (by NURBS [22,23] or finite element meshes [24,25]), and to determine the displacement field to apply to minimize the correlation residual computed over the entire region of interest. This technique provides a dense description of the measured fields (i.e., 3D shape or 3D displacement) and ensures direct links with current CAD and isogeometric softwares [26].…”

“…The minimization problem includes an extremely small number of unknowns, but cannot represent surfaces having low order continuities (i.e., parts with groove or machining defect). Dubreuil et al [24] extended the previous approach to finite element meshes by deforming the surface along the local normal vector. Further, regularization strategies were proposed either by penalizing the Hessian of the minimization scheme [23] or by adding mechanical constrains [25].…”

“…This system will rely on global stereocorrelation. Unlike previous works [27,24,25], the surface will be deformed using displacement fields that represent global (and no longer local) defects. Very recently, Jailin [28] proved that it was possible to determine the amplitude of 8 vibrational modes using global DIC.…”