Today, computer tomography data (CT data) are not only used for materials testing but also for analyzing measurement data from components. To do this, in order to obtain the full surface information from an object, boundary surfaces have to be calculated from the three-dimensional CT model. The surface data is available either in the form of triangle meshes (STL files) or as point clouds. To create a facet, the Marching Cubes algorithm is commonly applied, which uses global threshold values to determine the boundary between air and material.
To apply the method in practice, Fraunhofer IPA has implemented diverse optimization measures, such as the elimination of ambiguities as well as efficient calculation processes and intelligent online methods to reduce the size of data. These make is possible to deal with the quantities of data generated and manage the necessary calculation and storage requirements. However, global threshold value methods have some disadvantages because the optimum threshold value may vary within a volumetric shape even if the material is identical.
In addition to local threshold value methods, the department “Machine Vision and Signal Processing” has also developed an adaptive surface extraction method based on active contours. This new method enables different materials to be separated and also takes different areas into account when determining surfaces from CT data. In future, a user will therefore be able to generate triangle meshes according to regions in the desired resolution, thus ensuring that surfaces are generated in a controlled manner.