The algorithm is fully tested and implemented for regular use at The Australian National University micro-CT facility for both circular and helical trajectories. It can in principle be applied to more general imaging geometries and modalities. It is as robust as manual alignment but more precise since we have quantified the effect of misalignment.

This paper reports on recent advances at the micro-computed tomography facility at the Australian National University. Since 2000 this facility has been a significant centre for developments in imaging hardware and associated software for image reconstruction, image analysis and image-based modelling. In 2010 a new instrument was constructed that utilises theoretically-exact image reconstruction based on helical scanning trajectories, allowing higher cone angles and thus better utilisation of the available X-ray flux. We discuss the technical hurdles that needed to be overcome to allow imaging with cone angles in excess of 60°. We also present dynamic tomography algorithms that enable the changes between one moment and the next to be reconstructed from a sparse set of projections, allowing higher speed imaging of time-varying samples. Researchers at the facility have also created a sizeable distributed-memory image analysis toolkit with capabilities ranging from tomographic image reconstruction to 3D shape characterisation. We show results from image registration and present some of the new imaging and experimental techniques that it enables. Finally, we discuss the crucial question of image segmentation and evaluate some recently proposed techniques for automated segmentation.

A practical experimental procedure for transmission X-ray ghost imaging (XGI) using synchrotron light is presented. The authors demonstrate the method, discuss data acquisition and analysis, and measure the point-spread function of an XGI system. The generalization of the methods for future experiments is also discussed.

Autofocus-corrected, theoretically-exact cone-beam reconstruction is a viable option for reducing acquisition time in high-resolution micro-CT imaging. It also opens up the possibility of efficiently imaging long objects.

Ghost tomography using single-pixel detection extends the emerging field of ghost imaging to three dimensions, with the use of penetrating radiation. In this work, a series of spatially random x-ray intensity patterns is used to illuminate a specimen in various tomographic angular orientations with only the total transmitted intensity being recorded by a single-pixel camera (or bucket detector). The set of zero-dimensional intensity readings, combined with knowledge of the corresponding two-dimensional illuminating patterns and specimen orientations, is sufficient for three-dimensional reconstruction of the specimen. The experimental demonstration of ghost tomography is presented here using synchrotron hard x-rays. This result expands the scope of ghost imaging to encompass volumetric imaging ( i.e. , tomography), of optically opaque objects using penetrating radiation. For hard x-rays, ghost tomography has the potential to decouple image quality from dose rate as well as image resolution from detector performance.

We present "dynamic tomography" algorithms that allow for the high-resolution, time-resolved imaging of dynamic (i.e., continuously time evolving) complex systems at existing x-ray micro-CT facilities. The behavior of complex systems is constrained by the underlying physics. By exploiting a priori knowledge of the geometry of the physical process being studied to allow the use of sophisticated iterative reconstruction techniques that incorporate constraints, we improve on current frame rates by at least an order of magnitude. This allows time-resolved imaging of previously intractable processes, such as two-phase fluid flow. We present reconstructions from experimental data collected at the Australian National University x-ray micro-CT facility.

No abstract

Here we present an experimental and numerical investigation on the grain-scale geometrical and mechanical properties of partially crystallized structures made of macroscopic frictional grains. Crystallization is inevitable in arrangements of monosized hard spheres with packing densities exceeding Bernal's limiting density ϕ(Bernal)≈0.64. We study packings of monosized hard spheres whose density spans over a wide range (0.59<ϕ<0.72). These experiments harness x-ray computed tomography, three-dimensional image analysis, and numerical simulations to access precisely the geometry and the 3D structure of internal forces within the sphere packings. We show that clear geometrical transitions coincide with modifications of the mechanical backbone of the packing both at the grain and global scale. Notably, two transitions are identified at ϕ(Bernal)≈0.64 and ϕ(c)≈0.68. These results provide insights on how geometrical and mechanical features at the grain scale conspire to yield partially crystallized structures that are mechanically stable.