X-ray micro-tomography is a non-destructive technique for high-resolution 3D imaging of specimens of interest in research fields such as material science, (e.g. foams, granular packings), medicine, biology, paleontology, and geology. It is not a direct imaging technique but rather the 3D image is "computed" from a set of 2D X-ray images of the specimen taken from various directions.
The micro-tomography facility at ANU uses a broad spectrum of X-ray energies for imaging (a.k.a. polychromatic radiation). The attenuation properties of materials are a function of X-ray energy; low energy (or soft) X-rays are preferentially attenuated causing the x-ray beam to contain a greater proportion of high energy (or hard) X-rays as it passes through the specimen. This process is termed "beam hardening" and causes the set of 2D X-ray images to be inconsistent and leads to a computed 3D image with cupping, streaking, and halo artifacts that degrade image quality. It is the high-density objects in specimens in particular, (e.g., minerals in rocks, metal-pins in biological specimens), that cause significant streaking and halo artifacts (and in some cases make the final 3D image unusable).
A hardware solution is to "pre-harden" the X-ray beam, i.e., filter the X-ray source. This can mitigate the effect, however, there is a loss of sensitivity to components of lower attenuation. Several software techniques also exists to cope with high-density objects; this project will explore the performance of these methods as well as investigate alternatives with the aim to improve on the image quality resulting from "pre-filtering."
Willingness to engage with mathematical, physical, and computational disciplines. Familiarity with one or more of python/c/c++ is preferable.