A new method of performance verification for x-ray computed tomography measurements

Abstract

Through its ability to measure otherwise inaccessible internal structures, x-ray computed tomography (CT) has become a useful tool in the field of dimensional metrology. However, the lack of international standards for performance verification and the determination of metrological characteristics has so far prevented it from becoming a standardized measurement technology. This paper considers how a test object comprising a dismantleable tetrahedral assembly of four calibrated alumina spheres can be employed to assess the metrological performance of an industrial-type CT system. The measurement accuracy was characterized in terms of sphere size, separation and form errors. In all cases, sub-voxel accuracy was achieved, with errors as small as 1/10 of a voxel being obtained. Good measurement repeatability was demonstrated even upon dismantling and reconstructing the reference object. The tetrahedral test object offers a simple robust solution for the verification of metrological performance with great versatility and adaptability.

Towards in-process x-ray CT for dimensional metrology

“…It is noteworthy that all measurements in this section are sub-voxel in accuracy. This is to be expected and has been shown by a number of previous authors [9,27,[30][31][32] that the method of voxel scaling can frequently result in such an occurrence. The Nikon measurements are naturally the most accurate given the resultant voxel size is nearly 7.5 times smaller, but this does not necessarily equate to a proportional increase in measurement accuracy due to other source of error that can occur within x-ray CT.…”

“…Even with the number of potential errors, the dimensional measurement performance of CT systems has been shown to be of the order of 0.1-0.3 times the resolution of the system itself [9,28,[29][30][31][32]. To achieve such comparable errors the feature must be observable at the resolution of the system, and normally a form of correction strategy is applied to the acquired volume.…”

X‐ray computed tomography and its potential in ecological research: A review of studies and optimization of specimen preparation

“…Besides, techniques for calibration are been developed, which would greatly increase the accuracy of the measurements (Léonard, Brown, Withers, Mummery, & McCarthy, 2014). …”

Impact of the Threshold on the Performance Verification of Computerized Tomography Scanners

“…• Müller et al [20] propose the measurement of a simple ruby ball plate, which can be calibrated by means of a coordinate measuring machine; • Welkenhuyzen et al [21] studied in particular the problem of the verification of an high voltage CT scanner by means of a "forest of styli" as reference artifact; • a simple artifact constituted by four alumina balls shaped as a tetrahedron is proposed by Léonard et al [22] as reference artifact. The authors claim that "a sub-voxel accuracy was achieved with errors as small as 1/10 of a voxel obtained for the size error".…”

“…However, performance verification should be always performed according to some procedure recognized in international standards [22], but these standards have not been published yet, and the discussion on them is still ongoing [23].…”