Premium
Markings for Image‐Based Deformation Measurement on a Torsion Test Machine
Author(s) -
Joseph S. H.
Publication year - 2009
Publication title -
strain
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.477
H-Index - 47
eISSN - 1475-1305
pISSN - 0039-2103
DOI - 10.1111/j.1475-1305.2008.00425.x
Subject(s) - torsion (gastropod) , displacement (psychology) , contrast (vision) , displacement field , mathematics , deformation (meteorology) , digital image correlation , enhanced data rates for gsm evolution , artificial intelligence , optics , materials science , computer science , structural engineering , engineering , physics , composite material , finite element method , medicine , psychology , surgery , psychotherapist
The markings used for full‐field strain measurement by image correlation often take the form of a spray from aerosol or brush. On commercial test machines, specimens are more usually marked with an arrangement of dots or lines. Factors influencing the choice of markings for a specimen for a torsion machine are investigated. A theoretical analysis for predicting the random errors in displacement estimates of an image patch containing a mark using a sum‐squared difference criterion is developed, and compared with test results on synthetic images of edge and dot models. Tests on real images containing a contrast gradient are carried out at finite displacements, employing various criteria. The scatter of displacement estimates is compared with the random error predicted by the theoretical analysis, and the contrast gradients are measured. Further tests of the dot model on synthetic images quantify the systematic errors in displacement estimate that arise in the presence of contrast gradients. Further tests of dot and spray markings in synthetically sheared images indicate the extent to which displacement gradient parameters may be omitted from the criteria without increasing the random error of displacement estimates for deformed specimens.