Composite parts with 3D-textile reinforcement show promise in high-performance applications. For widespread use, accurate material characterisations are required. Characterisation of the textile architecture in the as-manufactured state may be performed with X-ray CT. Due to the similarity between the chemical composition of carbon fibres and epoxy based matrices, the contrast of X-Ray CT scans is poor. Therefore, segmentation with classical methods is difficult or even impossible. Alternatively, machine learning based segmentation approaches may be used. One drawback of machine learning-based algorithms is the need for datasets whose ground truth labellings require extensive manual labour. This can be circumvented by utilising automatically labelled synthetic X-ray CT data. In this study, a novel pipeline that generates synthetic CT image datasets, with automatically labelled ground truths, is developed. The pipeline is entirely based on free and/or open source software. As a segmentation model, we need to train a 3D U-Net with labelled volumes of considerable size. So far, initial studies based on training limited by the performance of a high-end personal computer shows promise, but we believe that by increasing the volume of each 3D data point, we can achieve even higher level of accuracy of the final model. As such, we need to train the model on hardware where we have at least ~40 GB or RAM.