Abstract
We introduce a beam-hardening correction method for lab-based X-ray computed tomography (CT) by modifying existing iterative tomographic reconstruction algorithms. Our method simplifies the standard Alvarez–Macovski X-ray attenuation model [Phys. Med. Biol. 21, 733 (1976)] and is compatible with conventional (i.e., single-spectrum) CT scans. The sole modification involves a polychromatic projection operation, which is equivalent to applying a weighting that more closely matches the attenuation of polychromatic X-rays. Practicality is a priority, so we only require information about the X-ray spectrum and some constants relating to material properties. No other changes to the experimental setup or the iterative algorithms are necessary. Using reconstructions of simulations and several large experimental datasets, we show that this method is able to remove or reduce cupping, streaking, and other artefacts from X-ray beam hardening and improve the self-consistency of projected attenuation in CT. When the assumptions made in the simplifications are valid, the reconstructed tomogram can even be quantitative.
© 2020 Optical Society of America
Full Article | PDF ArticleMore Like This
Nathanaël Six, Jan De Beenhouwer, and Jan Sijbers
Opt. Express 27(23) 33670-33682 (2019)
Amogha Pandeshwar, Matias Kagias, Zhentian Wang, and Marco Stampanoni
Opt. Express 28(13) 19187-19204 (2020)
Aimin Yan, Xizeng Wu, and Hong Liu
Opt. Express 25(20) 24690-24704 (2017)