Sixth GAMM Seminar on Microstructures - Abstract
Rumpf, Martin
Nowadays image acquisition in materials science allows the resolution of detailed microstructures in materials. Examples are microscopic pattern with interfaces of certain orientation or lattice structures where atoms are represented by single image dots. From these images macroscopic material parameters such as interface orientation, angles formed by segments of the interface, or the geometry of grain boundaries has to be extracted in a reliable way. Here, variational methods in image processing enable an automatic identification of such parameters which are described on the microscale and might vary macroscopicly. In particular, an anisotropic total variation type functional will be introduced which allows the identification of edge orientations and angles between edges at corners on the microscale. Furthermore, a Mumford Shah type model for the segmentation of grain boundaries and liquid solid interfaces on an atomic scale will be discussed.