Optimised region-of-interest-reconstruction in computed tomography and laminography using a-priori information

In non-destructive testing it is often required to investigate only a certainpart of the object, a region-of-interest. For computed tomography or computed laminography these kinds of reconstructions usually suffer from severe artefacts. Although there exists an approach which can be used to reduce these artefacts,it imposes a tight restriction on the form of the object, as it only works forplanar objects. In this work this approach is taken up and adapted in order to be able to cope with arbitrarily shaped objects. Therefore a-priori volumes are used which giveadditional information about the geometrical contour of the examined object.Using a-priori information leads to the two main tasks of this work: The first task is to integrate the a-priori information into the existing algorithms in orderto calculate the correction necessary for region-of-interest reconstructions. Thesecond task is caused by the file sizes of such a-priori volumes. As those needed for region-of-interest reconstructions can be quite large, an approach is examinedwhich allows to use down sampled a-priori data, namely linear interpolation and cubic spline interpolation.