Project Details
Projekt
Metrologically robust, massive acceleration of industrial computed tomography based on artificial intelligence.
Subject Area
Measurement Systems
Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
Production Systems, Operations Management, Quality Management and Factory Planning
Image and Language Processing, Computer Graphics and Visualisation, Human Computer Interaction, Ubiquitous and Wearable Computing
Production Systems, Operations Management, Quality Management and Factory Planning
Term
since 2023
Project identifier
Deutsche Forschungsgemeinschaft (DFG) - Project number 513552520
The aim of the proposal is to massively accelerate industrial high-quality CT measurements from the up-to-date hourly to the minute range (Inline CT range) while simultaneously ensuring metrological quality. The quality of the resulting CT measurements should significantly increase compared to the regular procedure of Inline CT, which is based on the empirical selection of the measuring parameters such as tube current, exposure time, and SOD and leads to quality losses. For this purpose, the attribute-specific, physical correlation is to be described by mathematical models within the project's scope to determine the attribute-related measurement uncertainty as a function of the given CT parameter set. Since the high number of projection images primarily defines the acquisition time for high-quality CT measurements, a significant reduction of projection images is targeted, which would lead to artefacts in the reconstructed volumes without further corrections. Therefore, neural network-based artefact correction methods are developed within the proposal, combining or modifying the projections (sinogram-domain) and the 3D volume data (image-domain). For the 100% inspection in the production process (inline CT), the inspection of a series of components is to be considered, i.e. the inspection of identical components, as they can be found, e.g. in a production line. It is assumed that the initially longer time for identifying suitable CT acquisition parameters is justified to significantly accelerate the subsequent permanent inline inspection of the component series.
DFG Programme
Research Grants
