Article first published online: 11 NOV 2008

Strain Volume 44, Issue 6, pages 421–422, December 2008

There are many instances in science where a good idea sits largely unexploited for a number of years and then suddenly triggers a rapid burst of research and development activity. Sometimes this is because a new field suited to its application opens up; in other cases new instrumentation or improvements in computing speed turn the method from a laborious one into a fast and convenient practical tool. X-ray tomography is a good example where the development of CCD detectors capable of recording digital images combined with improvements in computer power suddenly made three-dimensional reconstruction possible.

In the area of strain measurement, the concept of correlating two successive images in order to deduce the velocity, displacement or strain field is well established, having being exploited at least as long ago as the 1960s [1] to record the velocity of a moving image for aerial camera systems. Very little development work was carried out until the early 1980s when the technique was applied to measure strain [2]. At that time the correlation of a 15 × 15 pixel patch to find the most probable registration in terms of displacement, rotation and strain over an array of 121 points would take some 100 s. A sub-pixel accuracy of 0.2 pixel was regarded as ambitious. Perhaps due to poor displacement resolution and slow correlation times, few papers were forthcoming in the remainder of the 20th century, although a number of companies offering commercial image correlation software emerged.

Since then the growth of the topic has been dramatic both in terms of the development of the method and the diversity of its application. Strain reflects this trend with many articles currently in process or recently in print including a particularly helpful introduction to the analytical side of the topic by Hild and Roux [3].

The appeal of the method lies in the simplicity of the approach; images for analysis can be acquired across a very wide range of scales by video camera, optical microscope, confocal microscopy, scanning electron microscope, atomic force microscope or scanning tunnelling microscope. A high level of fine scale contrast is needed in the images in order to accurately correlate the images. Sometimes this can be achieved simply by exploiting the contrast inherent in the image, be that due to scratches or contrast between the constituent phases [4]. Otherwise it can be conveniently applied by means of paint or by a controlled deposition process. The disadvantage of paints and coatings is that the scale of the speckle needs to be tailored to the spatial scale of the images and may obscure microstructural features that may be of interest. Achievable displacement resolution and patch size (spatial resolution) are inversely related but 0.05 pixel resolution can easily be achieved currently for a 32 × 32 pixel patch [3]. The approach can be used to determine surface strains in two dimensions using a single camera, or in three dimensions using two cameras [5] or even in three dimensions within solid objects using images acquired by X-ray tomography [6–8].

Digital image correlation (DIC) is finding particular application for samples not amenable to the fixing of strain gauges, for example foams, powders and biological samples. Indeed the strain has been measured in a whole range of biological samples from arteries to the mechanical properties of human cells [9] and the tympanic membrane [10]. While the appeal of the method is that it can be applied over scales ranging from insect wings [11] to aircraft wings [12] and from plate tectonics to nanotechnology, it is being used to study affine continuum effects [2] and microstructural phenomena relating to grain or phase heterogeneity [4, 13]. Strain measurements in micro- and nanoscale devices are attracting particular attention [14]. Here the ability to machine, excavate and drill at the nanoscale by focused ion beam microscopy while at the same time imaging the resulting deflections by ion beam or field emission gun scanning electron beam is extending many of the traditional destructive techniques, such as slitting, hole drilling and cantilever deflection to the nanoscale [15, 16].

Another area of emerging interest is the use of DIC to investigate the state of strain local to a crack tip. This was one of the first applications of DIC allowing measurement of the mode I stress intensity factor (SIF) [17]. More recently, algorithms have been developed for determining mixed mode SIFs [18, 19]. The advantage of these methods is that, for the surface at least, it is possible to infer the crack-tip stress intensity in cases where analytical solutions do not exist, the crack tip is corrugated, or residual stresses or closure mechanisms might operate. In all these cases it is not a simple matter to infer the crack-tip stress intensity simply from the applied load.

In practice images can be acquired by high-speed camera [20]. Two applications are of special mention: dynamic failure [21] and high cycle fatigue [22]. In the former images have been acquired at 225 000 frames per second in a burst of 32 images for a polymer edge cracked beam subjected to impact loading. This enabled the SIF to be followed over 140 μs from the point of impact. In the latter, a frame rate of 1000 frames per second has been used to study the evolution of SIF over 77 Hz high cycle fatigue delineating the variation of KI and KII at 13 points during each cycle.

In summary, after a slow start, DIC is now progressing very quickly both in terms of the technical development and in terms of the applications for which it is applied. I am sure that we will see many more applications of the technique, not just within the traditional experimental mechanics sphere. Already the scales and timescales covered are impressive. I look forward to following the progress and achievements over the coming years within Strain.