Inaccuracy compensation in X-ray computed tomography image reconstruction and XCT-based finite element models

3D characterization techniques and modeling methods provide deep insights into materials and their microstructure on all length scales. In this thesis a hybrid 3D image reconstruction approach developed to minimize the common imaging artifacts in HR-XCT to improve the image reconstruction quality. Subsequently, the reconstructed 3D data converted into finite element models to estimate mechanical properties of complex material systems.

---

This thesis provides a solution to mitigate the effect of four common imaging artifacts in high-resolution X-ray computed tomography which ultimately reduce the accurateness of the acquired 2D projection images and of the reconstructed 3D data of the materials and objects studied. A novel hybrid tomographic image reconstruction approach was developed and implemented for parallel-beam and cone-beam geometries. The proposed approach has proven to suppress image artefacts effectively as demonstrated in hierarchical MoNi₄/MoO₂@Ni electrocatalyst system, epoxy molding compound and Didymosphenia geminate frustule. Especially deep learning-based approaches excelled in motion compensation and data recovery while statistical minimization performed adequately for misalignment compensation. Consequently, the reconstructed accurate 3D data converted into FE models to simulate the mechanical behaviours of complex material systems.