FoilTrack: a package to increase strain-resolution by improved X-radiographic image processing

In high pressure multi-anvil experiments X-radiography is used to ascertain strain in deforming samples because the tooling prevents optical or other direct observations of the sample. The processing of these X-radiographic images to determine bulk sample strain is one of the limiting factors to making measurements closer to the strains and strain-rates that occur during mantle convection or the passage of seismic waves. Typically, sample deformation in these experiments is tracked by the displacement of high-contrast marker foils in X-radiographs. X-radiographs are treated individually or pairwise in a multi-step process that tracks the displacement of marker foils during experiments. Here I develop a new algorithm, FoilTrack, that treats all the X-radiographic observations in a single-step process, resulting in improved accuracy and consistency of length changes determined from X-radiographic images, as well as providing more realistic parameter uncertainty. The improvements are demonstrated using data from small-strain sinusoidal deformation experiments.