WP1: CT Geometry Characterisation

The aim of this work package is to develop traceable and validated methods for the absolute geometrical characterisation of cone beam and parallel beam CTs. This will include the correction of errors based on 9 degrees of freedom (DoF) using reference standards, traceable calibration methods, in situ metrology systems and thermal models for instrument geometry correction, as well as the correction of errors originating from the X‑ray source and the detector in order to improve the accuracy of CT. The work will primarily focus on cone beam CT systems due to their abundance in industry. In addition, parallel beam CTs, such as the synchrotron CT, are considered to be powerful reference tools for providing mono‑energetic radiation.

In Task 1.1, reference standards will be developed and calibrated to efficiently determine the geometry of cone beam and parallel beam CTs. The implementation of such methods on commercial CT systems will be straightforward, thus simplifying uptake by industry. The developed standards will be used in WP1 – WP5.

In Task 1.2, the temperature distribution of CT systems will be measured and modelled. This is especially important as CTs contain major heat sources, such as the X‑ray source and the detector. Based on the results, strategies to reduce and correct temperature induced CT geometry deviations will be developed.

In Task 1.3, X‑ray source, and detector specific, effects on the CT geometry will be investigated. Therefore, a novel in situ X‑ray source tracking system will be implemented and the penetration depth of the X‑rays into the detector will be characterised and corrected.

In Task 1.4, advanced methods to determine the CT geometry with an accuracy better than 5 µm will be developed and evaluated. In addition to the reference objects and methods developed in Task 1.1 and Task 1.3 further measurement systems, such as laser interferometers, will be employed. The combined results will be used as an input for the development of correction methods to improve the accuracy of CT data.

WP2:  Advanced CT Methods

The aim of this work package is to develop advanced and traceable methods for dimensional CT measurements.

In Task 2.1, the effects that occur when multi‑material objects are measured will be investigated. CT systems with X‑ray tube (cone beam) and synchrotron CT will be taken into consideration. A supplementary material characterisation via atomic number and density will be carried out using a synchrotron CT.

In Task 2.2 sculptured surfaces, also called freeforms, will be measured using CT. Traceability will be achieved using at least one calibrated reference standard.

In Task 2.3, roughness characterisation using high resolution XCT will be investigated, including optimised measurement parameters, the limitations and the appropriate surface parameters.

WP3:  Fast CT

In this work package, the aim is to improve the efficiency of CT scans will be improved. Scans using X‑ray projections from circular trajectories, with few, or limited angles will be investigated and the reconstruction algorithms for inline applications, based on sparse modelling will be studied with the aim of significantly reducing the scan and reconstruction time to a few minutes (rather than a few hours). Different modalities of CTs with robotic arms will also be investigated to suit wider industrial applications.

In Task 3.1, existing reconstruction algorithms, the potential of advanced reconstruction algorithms and the requirements of fast CT will be reviewed. The ability of the consortium’s existing CTs to be used for fast CT scanning will be surveyed and a benchmark strategy and procedures for testing reconstruction algorithms will be developed. Both simulation data and experimental data will be prepared for use in the tests.

Task 3.2 will improve the selected reconstruction algorithms and will develop a metrology software toolkit with the possibility of adding post‑processing to the reconstruction. The improved algorithms will be able to deal with sparse and noisy data. The task will also survey the computation facility, and X‑ray sources and detectors for fast CT.

Task 3.3 will characterise the improved reconstruction algorithms from Task 3.2 for use in measurement tasks with limited angle measurement, few projection measurements and noisy measurements. Both the speed and dimensional accuracy of the algorithms will be considered.

Task 3.4 will further investigate fast industrial CTs that use a non‑circular trajectory. This task will expand the investigation to include industrial CTs with robotic arms. The report on the measurement results is expected to provide guidance for future developments.

WP4:  Simulation-Based Uncertainty Evaluation

The aim of this work package is to develop traceable methods for uncertainty estimation using virtual CT models and Monte‑Carlo simulations. With the availability of software packages like aRTist and CIVA which can provide nearly realistic simulations of CT imaging systems, uncertainty estimation is, in principle, possible based on virtual CT models, but it is not regularly applied. A methodology to perform an uncertainty estimation of CT, via simulation, will go beyond the state of the art.

In Task 4.1, software tools will be developed to perform batch calculations of virtual CT models and automated data analysis of the results.

In Task 4.2, accurate virtual CT model parameters which strongly determine the quality of the simulation, will be determined through a systematic approach. This will be undertaken based on prior experience and with input and specimens from WP1.

In Task 4.3, correction methods for specific CT image forming artefacts will be developed which will help to improve the quality of dimensional CT measurements. Correction filters will be developed to improve the quality of the 2D projections before reconstruction, and the influence on the measurand and measurement uncertainty will be evaluated both experimentally and by simulation.

In Task 4.4, the improved software tools from Task 4.1 and the adjusted reference object models from Task 4.2, will be applied to launch large computer simulations, to estimate the uncertainty of specific real CT systems and for comparison with the experimentally determined uncertainty.