Micro-CT is widely used in materials science and engineering applications because of its advantage of imaging and analyzing materials non-destructively. However, various errors can occur during the scan process. Scientists at the CT Scanner Unit of the Central Analytical Facilities at Stellenbosch University evaluated micro-CT image quality and demonstrated a simplified, reliable image quality measurement method which can be implemented easily. One of the measures used is sensitive to contrast and noise while the other is sensitive to image blur. The combination of these quality measures are sensitive to all typical micro-CT scan errors and artifacts as demonstrated in the work published in Materials Today Communications.

Researchers Prof Anton du Plessis, Muofhe Tshibalanganda and Stephan le Roux decided on this study due to the fact that X-ray tomography is growing in popularity, but lacked a simple, widely-used image quality metric that can capture all important aspects of the quality of a scan. Often users and researchers are not even aware that image quality can vary significantly depending on scan time, for example. A large variety of instrument types are available, differences in hardware and software and different skill-sets and experience levels of scientists all have an effect on the eventual quality of the scan.

For this study a 10mm cube of titanium alloy was used. The method involves simple measurements from raw CT data. In the first step the values are obtained from the entire cube material data and all air data outside the cube. In the second step the same analysis is made using values obtained from the region near the edge of the cube only (air and material).  The first step is sensitive mainly to contrast and noise and the second is more sensitive to blur or loss of sharpness which occurs more at the edges of the material. It was shown how an increase in scan time correlates with higher image quality which transfers to an improved analysis reproducibility.

The implications of this study are that image quality measurements can be used to support micro-CT-based materials analysis – a minimum quality value can be provided for a reliable measure. This will assist in a wider understanding of the importance of obtaining high quality CT data and not focusing on lower-cost “fast" scan times, for advanced applications.

For more information contact:

Prof Anton du Plessis

Email: anton2@sun.ac.za