CC BY-NC 4.0Blarr, JulianeJulianeBlarrKresin, NoahNoahKresinKrauß, ConstantinConstantinKraußWeidenmann, Kay AndréKay AndréWeidenmannLiebig, Wilfried V.Wilfried V.LiebigElsner, PeterPeterElsner2023-03-172023-03-172022https://publica.fraunhofer.de/handle/publica/437775https://doi.org/10.24406/publica-106410.24406/publica-10642-s2.0-85149395820When investigating the mechanical behavior of fiber-reinforced polymers, fiber orientation plays a decisive role concerning anisotropy. Fiber orientation distributions are typically measured in the form of fiber orientation tensors. In order to measure orientation tensors, computed tomography scans and consecutive image processing methods have become one of the leading non-destructive testing methods. The conflict between scan resolution and sample size limits the volume that can be scanned. To obtain the fiber orientation behavior across an entire plate, a direct interpolation of orientation tensors computed from CT scans of smaller volumes at selected coordinates of the plate is implemented. Rather than a component-based interpolation, the authors chose a decomposition and reassembly method interpolating shape and orientation of the tensors separately. While this approach has been implemented and used for e.g. diffusion tensors in medical imaging, the authors consider the application to sparse but measured CT-based data to be a novelty.enimage processingnon-destructive testingquaternionstensor algebratensor mappingconference paper