Reduction of metal artifacts for industrial computed tomography using multipositional data fusion

Metal artifacts represent a major challenge for measuring with X-ray computed tomography. This paper presents a method for multipositional data fusion for the reduction of metal artifacts. For this purpose, several scans of the same object are performed with different object positioning, registered on each other and then reconstructed together using a local quality measure. The added value of the method is demonstrated in practical experiments. In particular, it is shown how plastic structures can be made visible despite strong metal artifacts with little effort and without prior knowledge, giving the method a unique selling point compared to existing metal artefact reduction methods.