An adaptive thresholding algorithm for porosity measurement of additively manufactured metal test samples via X-ray computed tomography
An adaptive thresholding algorithm for porosity measurement of additively manufactured metal test samples via X-ray computed tomography
If additively manufactured metal parts are to be used as performance critical components in the aerospace and other high-value industries, then non-destructive methods for measuring the porosity of these parts are required; X-ray computed tomography (XCT) is a promising instrument for this measurement task. One disadvantage of using XCT for this purpose is the presence of artefacts in XCT data of high-density materials (aluminium, titanium, steel and Inconel). These artefacts are problematic for segmentation algorithms that are required for porosity measurement. In this work we propose an adaptive thresholding algorithm for segmenting XCT data. The algorithm is specifically designed to overcome the influence of beam hardening (cupping) artefacts that are typically observed when XCT scanning high-density metal parts. The algorithm is also designed to have as few as possible user-defined settings in order to minimise the influence the user has on the segmentation result. The results show that the algorithm is able to segment small voids in low contrast-to-noise ratio XCT data and that the algorithm is robust to cupping artefacts. The accuracy of the method is evaluated for a set of Inconel cubes approximately 10 mm × 10 mm × 10 mm in size (H × W × D), reference porosity measurements are made using Archimedes’ principle. The mean error and standard deviation of the XCT percentage porosity measurements evaluated by four different users are found to be: −0.38 ± 0.09%, 0.09 ± 0.09% and −0.23 ± 0.22% for samples with bulk porosities of 0.23 ± 0.28%, 1.42 ± 0.28% and 3.00 ± 0.28%, respectively.
Lifton, Joseph
9be501ec-2742-4ab6-8a5a-996c5b7c23ae
15 February 2021
Lifton, Joseph
9be501ec-2742-4ab6-8a5a-996c5b7c23ae
Lifton, Joseph
(2021)
An adaptive thresholding algorithm for porosity measurement of additively manufactured metal test samples via X-ray computed tomography.
Additive Manufacturing, 39, [101899].
(doi:10.1016/j.addma.2021.101899).
Abstract
If additively manufactured metal parts are to be used as performance critical components in the aerospace and other high-value industries, then non-destructive methods for measuring the porosity of these parts are required; X-ray computed tomography (XCT) is a promising instrument for this measurement task. One disadvantage of using XCT for this purpose is the presence of artefacts in XCT data of high-density materials (aluminium, titanium, steel and Inconel). These artefacts are problematic for segmentation algorithms that are required for porosity measurement. In this work we propose an adaptive thresholding algorithm for segmenting XCT data. The algorithm is specifically designed to overcome the influence of beam hardening (cupping) artefacts that are typically observed when XCT scanning high-density metal parts. The algorithm is also designed to have as few as possible user-defined settings in order to minimise the influence the user has on the segmentation result. The results show that the algorithm is able to segment small voids in low contrast-to-noise ratio XCT data and that the algorithm is robust to cupping artefacts. The accuracy of the method is evaluated for a set of Inconel cubes approximately 10 mm × 10 mm × 10 mm in size (H × W × D), reference porosity measurements are made using Archimedes’ principle. The mean error and standard deviation of the XCT percentage porosity measurements evaluated by four different users are found to be: −0.38 ± 0.09%, 0.09 ± 0.09% and −0.23 ± 0.22% for samples with bulk porosities of 0.23 ± 0.28%, 1.42 ± 0.28% and 3.00 ± 0.28%, respectively.
This record has no associated files available for download.
More information
Published date: 15 February 2021
Identifiers
Local EPrints ID: 498007
URI: http://eprints.soton.ac.uk/id/eprint/498007
ISSN: 2214-8604
PURE UUID: 3efa84ae-66fa-4d08-90ea-44d9ab3a05cb
Catalogue record
Date deposited: 05 Feb 2025 18:19
Last modified: 06 Feb 2025 03:12
Export record
Altmetrics
Contributors
Author:
Joseph Lifton
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics