X-ray Computed Tomography (CT) is a highly valuable non-contact technology for performing non-destructive measurement tasks. However, analyzing data obtained from CT scans can be challenging due to the presence of image artifacts, which obscure scans and make product analysis and evaluation increasingly difficult.
Whilst image artifacts can be minimized using a variety of methods, they can be an unfortunate reality and it is therefore highly beneficial for CT analysts to appreciate the different forms they can take.
Image Credit: North Star Imaging
A CT image artifact is any divergence between the reconstructed values produced in a CT image and the actual sample values based on its material density and geometry. These artifacts are predictable in pattern, usually appearing in scans as bright/dark streaks or shadows.
Red arrows indicate examples of CT image artifacts
Causes of CT Image Artifacts
Primarily, artifacts can be caused by one of two ways; issues with the CT setup, e.g. ring artifacts, under-sampling and sample movement, or issues concerning the sample, e.g. beam hardening, scattered radiation and minimal X-ray penetration.
Artifacts caused by issues with the CT system do not tend to cause major problems as they can be removed with a correct scan technique. Sample-dependent artifacts, on the other hand, are significantly more challenging as their effects can be reduced (with CT techniques/software corrections) but are not always eliminated. The majority of sample-dependent artifacts are caused by a phenomenon called beam hardening.
X-ray photons coming from a tube are made up of a full spectrum of X-ray energies, not just the voltage that the tube is set at. The energy of a photon determines its level of attenuation when passing through a sample, with low-energy photons attenuating much faster than high-energy ones. Therefore, when an X-ray beam begins to penetrate a sample material, the lower energy X-rays preferentially attenuate, resulting in an overall higher energy of the X-ray beam - ‘beam hardening’.
The effect of beam hardening on a single material can usually be managed by adjusting CT reconstruction algorithms. However, if the sample contains multiple materials within a single scan volume, imaging differing densities simultaneously tends to result in artifacts.
Minimizing Beam Hardening Artifacts
In general, the harder the X-ray beam (i.e. the higher the average energy), the less beam hardening will impact the CT image. This can be achieved with use of a higher tube voltage or by adding copper or brass filters onto the face of the tube. Making sure that the densest regions of the sample do not overlap with each can also help, as can performing a helical scan instead of a standard cone beam scan. However, it must be taken into consideration that these methods can reduce image contrast which may impact the overall quality of the CT image.
Alternatively, a beam hardening correction factor can be applied as part of CT reconstruction software to further minimize artifacts. As these methods are sample dependent, some experimentation may be required to achieve the maximal outcome.
What to Do When Artifacts Cannot Be Fully Eliminated
Even the best CT technique and software corrections are not able to minimize every image artifact, resulting in interpretation of a non-ideal scan. When this happens, a good comprehension of the forms an artifact can take is necessary to reduce possible misinterpretation of sample defects.
Examples of beam hardening “shadow” artifacts along the edges of flat, dense surface boundaries. Note that artifacts are especially accentuated along long, flat surfaces versus rounded or irregular objects.
Even though there may be valuable information within the artifact, it is typically essential to window level the image to offer adequate viewing within specific regions of interest. Adjusting the histogram can allow the interpreter to determine whether an artifact is obscuring anything important. However, if the streaks or shadows produced by the artifact are very disruptive, entire regions of the sample can be declared un-interpretable. In such cases, it is impossible to confirm if there is a defect in an area of the sample, and so performing another CT scan or using a different analysis technique is recommended.
Image artifacts in a CT scan are common in samples with a combination of different materials or varying material thicknesses. Without a good understanding of what they look like or where they come from, artifacts can make sample analysis of CT scans highly challenging and increase the risk of image misinterpretation and false verification of products.
This information has been sourced, reviewed and adapted from materials provided by North Star Imaging, Inc.
For more information on this source, please visit North Star Imaging, Inc.