Development of Digital Image Correlation tool to determine displacements and strains in mechanical testing

Ellon Bernardes de Assis¹; Francisco Evangelista Junior¹

¹ Universidade de Brasília

doi:10.20906/CPS/CILAMCE2017-1207

Resumo

This paper intends to implement a computational tool using DIC (Digital Image Correlation) for determining in-plane displacement and strain of objects subjected to mechanical testing. DIC is a robust technique for measuring full field of displacements and strain from comparison between object images on undeformed and deformed state. The first step of the routine is dividing the image in small subsets that have their displacement estimated by a first order function. The Zero Normalized Sum of Squared Differences is used as correlation criteria for comparison of the subsets. The Newton-Raphson algorithm is used to minimize the correlation criteria and determine the unknown displacement parameters of each subset. The calculation order of the subsets is guided by the correlation coefficient as proposed by the Reliabity Guided DIC algorithm. This allows the calculation of objects with irregular borders. Finally, the Savitsky-Golay algorithm is used for smoothing of the strain field. The implemented routine is tested using images artificially deformed with digital image editors and real mechanical test images. The results of the implemented routine are compared to the artificial predetermined displacements and to results from other open source DIC.

Palavras-chave: digital image correlation; displacement/deformation measurement; full field measurement; non-contact measurement; mechanical testing; reliability guided DIC