ABSTRACT--A three-dimensional extension of twodimensional digital image correlation has been developed. The technique uses digital image volumes generated through high-resolution X-ray tomography of samples with microarchitectural detail, such as the trabecular bone tissue found within the skeleton. Image texture within the material is used for displacement field measurement by subvolume tracking. Strain fields are calculated from the displacement fields by gradient estimation techniques. Estimates of measurement precision were developed through correlation of repeat unloaded data sets for a simple sum-of-squares displacement-only correlation formulation. Displacement vector component errors were normally distributed, with a standard deviation of 0.035 voxels (1.22 Ixm). Strain tensor component errors were also normally distributed, with a standard deviation of approximately 0.0003. The method was applied to two samples taken from the thigh bone near the knee. Strains were effectively measured in both the elastic and postyield regimes of material behavior, and the spatial patterns showed clear relationships to the sample microarchitectures.KEY WORDS--Strain measurement, X-ray tomography, image correlation, trabecular bone Many advanced synthetic foams and composites, and virtually all naturally occurring materials, present tremendous challenges to both the experimental and analytical branches of mechanics. Analytical methods must account for anisotropy, inhomogeneity and even nonlinearity in the elastic range, complex yield criteria and postyield behavior that can include hardening, softening, collapse and rebound phenomena. Determination of the extensive sets of constitutive parameters required for these materials, and validation of analytical results, is often beyond the capabilities of standard experimental techniques.The same features that generate complex behavior within these materials, however, also provide the opportunity for a new type of experimental strain mapping. Mixtures of different component materials, hierarchical microarchitectures and large-scale inclusions and voids are the tools used by Original manuscript submitted: May 1, 1998. Final manuscript received: January 22, 1999 both nature and the engineer to tune material behavior over a broad range of loading conditions. These same features are, in many instances, imageable in three dimensions by modern tomographic techniques. For materials of this type, we have developed a technique for estimating the full continuum-level strain tensor in three dimensions throughout the interior of test samples by correlating imaging volumes of the samples in unloaded and loaded states.The new method, digital volume correlation, is an extension of standard digital image correlation, 1-5 a technique that has found numerous applications 6-8 and has been extensively characterized. 9-11 Digital volume correlation relies on the X-ray tomographic imaging of naturally occurring material texture within samples.X-ray-based approaches have been used in two dimensions fo...

No abstract

Digital volume correlation, an experimental method for volumetric strain measurement, has experienced a growth in technique development and application since its introduction in 1999. This has largely been the result of more accessible volumetric imaging methods and greater speed and capacity of computational facilities. This paper reviews recent work using the method and presents examples from the author's laboratory. The emerging message is that, although closely related to the surface-oriented method of digital image correlation, the volume method is distinct due to its predominant reliance on naturally occurring image texture for displacement tracking. This requires careful tuning for successful application with different materials, and therefore the appropriate focus should not be on developing the 'best' digital volume correlation method, but on developing a set of tools that can be selected from and adjusted to specific mechanics problems.

Under conditions of maximal instability with similar material properties between specimens, differences in stiffness of posterior pelvic ring fixation can be demonstrated. The choice of which method to use is multifactorial.