Outstanding Paper Awards 2012

This paper presents physical and metrological characterization measurements conducted for an industrial x-ray micro-computed tomography (CT) system. As is well known in CT metrology, many factors, e.g., in the scanning and reconstruction process, the image processing, and the 3D data evaluation, influence the dimensional measurement properties of the system as a whole. Therefore, it is important to know what leads to, and what are the consequences of, e.g., a geometrical misalignment of the scanner system, image unsharpness (blurring), or noise or image artefacts. In our study, the two main components of a CT scanner, i.e. the x-ray tube and the flat-panel detector, are characterized. The contrast and noise transfer property of the scanner is obtained using image-processing methods based on linear systems theory. A long-term temperature measurement in the scanner cabinet has been carried out. The dimensional measurement property has been quantified by using a calibrated ball-bar and uncertainty budgeting. Information about the performance of a CT scanner system in terms of contrast and noise transmission and sources of geometrical errors will help plan CT scans more efficiently. In particular, it will minimize the user's influence by a systematic line of action, taking into account the physical and technical limitations and influences on dimensional measurements.

Particle image velocimetry (PIV) selects the maximum of the cross-correlation map to represent the modal displacement, and a wealth of information stored in the cross-correlation is discarded. We introduce a novel method, termed polynomial element velocimetry (PEV), which results in continuous velocity and velocity gradient measurements. PEV utilizes the extra information stored in the cross-correlation to determine continuous velocity measurements with low levels of measurement noise. In contrast to PIV, the continuous nature of velocity measurements facilitates the direct determination of the velocity gradient. The PEV method is applied to two laboratory flows: flow in a channel and flow behind a circular cylinder at Reynolds number, Re = 30, and is shown to greatly reduce the noise in the measurements. In addition, the accuracy of PEV is validated using two computer-generated synthetic flows: parabolic flow in a channel and flow past a square cylinder at Re = 30. In these cases, PEV is shown to reduce the velocity measurement error by up to 45% and the vorticity estimation error by up to 77% when compared to PIV. A key benefit of the PEV method is that it is capable of calculating continuous measures for flow gradient with greatly reduced bias errors. In particular, PEV provides a more accurate measurement of the vorticity near interfaces such as a cylinder wall or channel wall where PIV methods only provide measurement data at half the sampling window size from the wall. Since PEV utilizes the entire shape of the cross-correlation map to determine a local map for the underlying velocity, minimal random error is transmitted to the estimated flow gradient. This feature of the PEV method makes it optimal for flows where flow gradients are well defined and there are insufficient pixels to fully resolve structures in the flow using PIV.