[en] Purpose: In this paper we show that optimization-based autofocus may be used to overcome the instabilities that have, until now, made high-resolution theoretically-exact tomographic reconstruction impractical. To our knowledge, this represents the first successful use of theoretically-exact reconstruction in helical micro computed tomography (micro-CT) imaging. We show that autofocus-corrected, theoretically-exact helical CT is a viable option for high-resolution micro-CT imaging at high cone-angles (∼50 deg.). The elevated cone-angle enables better utilization of the available X-ray flux and therefore shorter image acquisition time than conventional micro-CT systems. Methods: By using the theoretically-exact Katsevich 1PI inversion formula, we are not restricted to a low-cone-angle regime; we can in theory obtain artefact-free reconstructions from projection data acquired at arbitrary high cone-angles. However, this reconstruction method is sensitive to misalignments in the tomographic data, which result in geometric distortion and streaking artefacts. We use a parametric model to quantify the deviation between the actual acquisition trajectory and an ideal helix, and use an autofocus method to estimate the relevant parameters. We define optimal units for each parameter, and use these to ensure consistent alignment accuracy across different cone-angles and different magnification factors. The tomographic image is obtained from a set of virtual projections in which software correction for hardware misalignment has been applied. Results: We make significant modifications to the autofocus method that allow this method to be used in helical micro-CT reconstruction, and show that these developments enable theoretically-exact reconstruction from experimental data using the Katsevich 1PI (K1PI) inversion formula. We further demonstrate how autofocus-corrected, theoretically-exact helical CT reduces the image acquisition time by an order of magnitude compared to conventional circular scan micro-CT. Conclusions: 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.

[en] Purpose: The authors present a robust algorithm that removes the blurring and double-edge artifacts in high-resolution computed tomography (CT) images that are caused by misaligned scanner components. This alleviates the time-consuming process of physically aligning hardware, which is of particular benefit if components are moved or swapped frequently. Methods: The proposed method uses the experimental data itself for calibration. A parameterized model of the scanner geometry is constructed and the parameters are varied until the sharpest 3D reconstruction is found. The concept is similar to passive auto-focus algorithms of digital optical instruments. The parameters are used to remap the projection data from the physical detector to a virtual aligned detector. This is followed by a standard reconstruction algorithm, namely the Feldkamp algorithm. Feldkamp et al.[J. Opt. Soc. Am. A 1, 612-619 (1984)]. Results: An example implementation is given for a rabbit liver specimen that was collected with a circular trajectory. The optimal parameters were determined in less computation time than that for a full reconstruction. The example serves to demonstrate that (a) sharpness is an appropriate measure for projection alignment, (b) our parameterization is sufficient to characterize misalignments for cone-beam CT, and (c) the procedure determines parameter values with sufficient precision to remove the associated artifacts. Conclusions: 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.

[en] Dry sand turns into a stiff and moldable material as soon as it is mixed with some liquid. This is a direct consequence of the internal liquid-air interfaces spanning between the grains which causes capillary cohesion by virtue of the surface tension of the liquid. As a model for wet granulates we investigated random packings of submillimeter spherical beads mixed with water. Measurements of the tensile strength and the fluidization threshold demonstrate that the mechanical stiffness is rather insensitive to the liquid content over a wide range. Only for a high liquid content, when more than half of the available pore space is filled with liquid, does the capillary cohesion weaken. In order to understand the interplay between the mechanical properties and the liquid content, we investigated the liquid distribution in random packings of glass spheres by means of x-ray microtomography. The three-dimensional images reveal that the liquid forms a network of capillary bridges fused at local triangular bead configurations. The spontaneous organization of the liquid into these ramified structures, which exhibit a large liquid-air interface, is responsible for the constancy of the cohesive forces in a wide range of liquid contents beyond the onset of capillary bridge coalescence.