Industrial Computed Tomography (CT) has recently gained a prominent role in the field of dimensional metrology as a
powerful 3D coordinate measurement technique. Its main advantage is the ability of measuring both inner and outer
features of geometrically complex workpieces without altering or damaging them. The settings with which CT data are
acquired can contribute to the uncertainty of measurements; these settings are chosen by the users in an intuitive way,
resulting in high variability of the measurement outcome. There is currently no available holistic model that can (1)
describe the relationship between CT setup parameters and measurement uncertainty, or (2) determine the optimal
parameters for a given measurement task. In this study, we propose an analytical method to optimize imaging parameters
for multimaterial measurements. The proposed method takes as input information about the nominal workpiece geometry
and material composition. For a given workpiece position and orientation, the optimal photon energy is calculated to
maximize edge detectability and minimize image noise. Subsequently, tube voltage and prefilter are chosen to ensure
that the generated X-ray spectrum is quasi-monochromatic around the optimal photon energy. Focal spot size is chosen
to minimize resolution, while also avoiding blurring. Then, the corresponding tube power is evaluated as well as the tube
current. Finally, integration time is minimized so that the gray values of air are still in the linear range of the detector.
The presented method was implemented into a software application and validated through simulation of CT scans of
multimaterial workpieces. Measurements with predicted parameters were found to be more accurate than those
performed with parameters chosen by expert users.
|