Skip to content

Latest commit

 

History

History
40 lines (34 loc) · 2.73 KB

paper.md

File metadata and controls

40 lines (34 loc) · 2.73 KB
title tags authors affiliations date bibliography
AXITOM: A Python package for reconstruction of axisymmetric tomograms acquired by a conical beam
Python
tomography
conical-beam
X-ray
name affiliation
Sindre Nordmark Olufsen
1
name index
Structural Impact Laboratory (SIMLab), Department of Structural Engineering, NTNU, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway
1
08 August 2019
paper.bib

Summary

The ability to picture the internals of a body in a non-invasive way is vital for both diagnostics and research in a broad range of fields. Tomography is the imaging of body sections, typically acquired from projections of the body by using a penetrating wave, and then reconstructed by employing a reconstruction algorithm. Some cases allow for simplification of the tomographic problem, one of them being the presence of symmetries within the tomogram. Axi-symmetry reduces the tomographic reconstruction to the inversion of the Abel transform, which can be obtained from a single projection, having a fractional acquisition cost compared to typical datasets. In home-laboratory X-ray setups, X-rays are distributed in space as a conical beam rather than parallel rays, which has to be accounted for in the reconstruction of the tomogram.

AXITOM is a Python package that allows for the reconstruction of axi-symmetric tomograms when the projections have been obtained by a conical beam. Excellent Python packages such as TomoPy [@Gursoy2014] and the Astra toolbox [@Pelt2016; @vanAarle:16; @VANAARLE201535] are already available for tomographic reconstruction, but are general-purpose and do not exploit axi-symmetries. PyAbel [@Hickstein2016] is a Python project for inversion of the Abel transform but is focused on parallel beam geometries.

AXITOM contains a collection of high-level functions that can be used to load datasets, perform virtual experiments and perform the tomographic reconstruction. A Feldkamp David Kress algorithm [@Feldkamp:84] is used to reconstruct the tomograms, with modifications for reduced computational cost. The implementation relies on NumPy [@Numpy] and SciPy [@SciPy], as well as numerous packages for visualization and IO.

AXITOM was implemented for reconstructing axi-symmetric density fields measured by X-ray absorption radiography. This project is a part of the ongoing research within the SFI CASA research group at NTNU and has been a key component in the pursuit of in-situ investigation of cavitation of polymers during deformation.

Acknowledgements

The author gratefully appreciates the financial support from the Research Council of Norway through the Centre for Advanced Structural Analysis, Project No. 237885 (SFI-CASA).

References