Techniques in helical scanning, dynamic imaging and image segmentation for improved quantitative analysis with X-ray micro-CT

Abstract

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.

The mandible and dentition of the Early Cretaceous monotreme Teinolophos trusleri

“…7) were produced using Micro-X-ray tomography on an instrument designed at the Research School of Physics and Engineering, Australian National University with 80kV/0.1mA tungsten filament-based X-ray source; the cone beam being pre-filtered with 1 mm aluminium. The projection series comprised 1500, 1024 × 1024 pixel images collected over 360°taking 4 h. In-house software was used for reconstruction (Sheppard et al 2014). Tomograms were visualized with the opensource volume renderer, Drishti, https://github.com/ AjayLimaye/drishti.…”

Fast laboratory-based micro-computed tomography for pore-scale research: Illustrative experiments and perspectives on the future

“…This means that the diameter of the flow cell should be small in order to achieve high magnifications and correspondingly high resolutions (depending on the characteristics of the scanner) while simultaneously maximizing the use of the available X-ray flux (i.e. without having to increase the focus-detector distance more than necessary) [1,16]. This is crucial to discriminate the pore network in the sample with sufficient detail.…”

“…Other complications which may occur are for example insufficient angular sampling, too long detector read-out times and limited rotation speed. Despite the fact that rather little attention has been given to this topic in the literature (with the exception of work performed by researchers at the Australian National University [14][15][16]), advances in micro-CT hardware and in reconstruction and analysis software are starting to render sub-minute pore-scale experiments possible at laboratory set-ups. In this work, we illustrate this by the visualization of drainage of a Bentheimer sandstone and convective/diffusive mass transport of a salt (CsCl) in a water-saturated Savonnières limestone.…”

Contribution for Content based Image Retrieval by Multiple Descriptors

“…However, due to limitations of the textual methods, development of methods based on visual content has become paramount. However, this explains by diverse application areas related to imaging science [1] [2] and the intense research activity devoted to CBIR (Content Based Image Retrieval) system in recent years. The objective of CBIR is automatic structuring image collections from selected visual criteria [3].…”