Data segmentation using computed tomography for the automotive industry and logistics

When it comes to preparing an inventory of entire vehicles, the automatic segmentation of three-dimensional datasets from X-ray computed tomography (CT) remains an unresolved challenge. Classical methods are unable to separate different parts and components into voxels and identify them with sufficient reliability. At present, this virtual »dismantling« process can only be performed manually, which is enormously expensive and time-consuming for the industry. There is, however, a great deal of interest on the part of industry, and strong demand for corresponding CT measurements there is an urgent need for solutions that can automatically break the data down into subgroups and convert the resulting volume images of individual assemblies into CAD-compatible formats.

Self-learning cognitive software systems in the area of machine learning (ML) – in particular those using convolutional neural networks (CNNs) – are regarded as the most promising technique for solving this problem. Systems of this kind should be able to automatically identify and segment image content (structures or components) in a complex dataset and characterize the content’s properties (position, size, shape). They should also be resistant to random variations in image quality or depicted objects and ensure flexible adaptation to gradual changes in components (e.g., varying size, shape, structure, or geometry).