Quality metrics for measurement uncertainty affected by the number of projections in xCT-based dimensional metrology

X-ray Computed Tomography (xCT) stands out as a Non-destructive Evaluation (NDE) method, enabling the simultaneous measurement of both internal and external component features with a single scan. From a metrological perspective, high-quality xCT measurements have low measurement uncertainties but require a long scanning time and an increased cost. This is primarily due to the required high number of projections obtained during the scan, since reducing the number of projections lowers the measurement quality and accuracy of the dimensional information acquired from the measurement. As a result, the broader integration of xCT systems in certain industrial applications is challenging. This work investigates the impact of varying the number of projections obtained during an xCT scan on different general-purpose quality metrics that are applied to the grey-level distribution of the reconstructed volume and correlate it with statistical analysis of extracted geometrical features of real xCT scans with different number of projections. A key focus of the work is comparing how different quality metrics respond to reductions in the number of projections acquired during xCT scans. Through this comparative analysis, the work seeks to identify the most effective metrics for indicating increased measurement uncertainty. The findings are expected to enable the calibration of the number of projections to balance precision, cost and time effectiveness, and efficiency in quality control for industrial applications.