Cloud-CT

Improved CT Image Reconstruction with AI Support

The increasing complexity in the automotive industry requires new quality assurance procedures. Driving factors include CO2 reduction, functional integration, lightweight construction, new manufacturing processes (e.g., gigacasting, hybrid processes, additive manufacturing), and the avoidance of costly recalls. In all cases, the requirements for non-destructive testing (NDT) are increasing. Computed tomography (CT) is particularly suitable for this purpose due to its material and structural resolution. However, current limitations relate to scan time, image quality, degree of automation, and IT integration.

The aim of the Cloud-CT project is to open up new possibilities for industrial CT through the use of AI-supported reconstruction methods. The focus is on reduced measurement times through data-reduced scans, improved image quality (including GANs and energy-based priors), modular software architecture for flexible applications, and cloud-based further processing through AI-supported data compression. The overall system is designed to enable automated, robust quality assessment of large and complex structural components such as gearboxes or battery cells, while simultaneously reducing costs and increasing throughput.

Project Partners

• Microvista GmbH, Blankenburg