Automated evaluation of the surface point quality in dimensional X-ray computed tomography

Abstract

X-ray computed tomography (CT) enables dimensional measurements of numerous measurands with a single scan, including the measurement of inner structures. However, measurement artefacts complicate the applicability of the technology in some cases. This paper presents a methodology to assess the surface point quality of computed tomography measurements without the requirement of a CAD model. Measurement artefacts lowering the surface point quality can therefore automatically be detected. The correlation of quality values with the random measurement error is demonstrated. The presented method can in principle be used to weight single fit points to reduce the measurement uncertainty of CT measurements.

Review of the influence of noise in X-ray computed tomography measurement uncertainty

“…However, the majority of research has focused on methods for measuring and decreasing noise, and there has been little research regarding the contribution of noise to measurement uncertainty. Although the volume of research is small, some researchers have studied how noise affects uncertainty of different type of measurements and proposed different ways to estimate its contribution [21,50,51,55,117,121,124].…”

“…beam hardening, scattering and blurring). The surface point quality metric developed by Muller et al [121] or the method to evaluate measurement uncertainty due to the ISO-50 surface determination algorithm proposed by Lifton et al [124] are examples of this trend. Due to the complexity of the relationship between noise and uncertainty of XCT measurements, we consider that such approaches are more suitable for addressing the assessment of noise contribution to XCT uncertainty than methods that attempt to estimate noise contribution alone.…”

“…Muller et al [121] developed a method to assess the surface point quality (SPQ) of XCT measurements without requiring a CAD model. The basic principle for this method was previously published elsewhere [122,123].…”

“…Additionally, a qualitative correlation between the SPQ and the random measurement error was demonstrated. These authors determined that the main challenges are the abstract nature of the studied quality parameters and the arbitrary scale of the SPQ values, which does not allow for comparability between measurements [121].…”

Comparison of Geometrically Derived Quality Criteria regarding Optimal Workpiece Orientation for Computed Tomography Measurements

“…However, the magnitude of deviations can not be derived from the a priori analysis. The results were additionally compared with an analysis of the single point quality (SPQ) of the determined surface according to [17]. Using this approach, grey value curves are extracted for points on the surface, using the direction of the local surface normal vector.…”