Development of a Computer Vision Methodology for the Analysis of Surface Deformations in Magnified Images

An optical system, similar in concept to a telecentric lens system, has been developed to decouple in-plane surface deformations from out-of-plane motions for magnified images of a specimen under loading. The key parameters of the optical system were identified and rigorous testing of a prototype system was completed. The range in system parameters that can be used to eliminate the effect of out-of-plane motion (ΔS_o) when ΔS_o is less than a pre-specified magnitude (β) was experimentally quantified. Using the test results, an optical system was designed, built and successfully tested.

Images of two test specimens at various loads are obtained at magnifications ranging from 2X to 6X, after testing was performed to ensure that the specimens do not move out-of-plane more than β during the loading history. First, images of a specimen undergoing uniaxial loading are acquired. Next, images of a single edge cracked specimen under load are obtained. For both specimens, the in-plane surface displacement field is determined by using digital image correlation of image subsets. Displacement gradients on the surface are computed after smoothing the displacement data. It is shown that the optical system developed in this work can be combined with a digital correlation method to accurately quantify in-plane surface deformations independent of the out-of-plane motion for magnifications up to 6X. Similar considerations will allow one to use the methodology with larger or smaller magnification factors.