Abstract:
© 2019 IEEE. Recent advances in digital image correlation (DIC) techniques now enable full-field measurements of surface shape and deformation using relatively inexpensive camera systems. Yet, established methods are not without limitations, exhibiting only modest accuracy and considerable sensitivity to noise effects. In order to address this, we have developed an advanced computer vision toolbox with which it is possible to accurately detect three-dimensional (3D) surface displacements to a practical resolution of 10 \mu \mathrm{m} over a field of view of approximately \rm{150 mm \times 150 mm}. By taking advantage of a high-resolution subpixel image registration algorithm and a stereoscopic imaging system, a methodological basis for a highly accurate and comprehensive deformation analysis is presented. This was demonstrated by the application of known forces to a simple cantilever structure while measuring resultant surface deformations. Measurements were then compared with expected displacement values predicted by analytical calculations as well as a finite element model. Comparisons showed our measurements were within 120 \mu \mathrm{m} of model predictions. Subsequently, we propose a method of non-contact force-sensing by exploiting the relationship between known applied forces and acquired displacement data. For our dataset, the system was able to predict force to within 0.38 %.
