Analiza Metod Detekcji Dyfrakcyjnych Linii Kikuchiego

The goal of the dissertation is to investigate and propose new methods for automatic Kikuchi lines detection. New subdivision of microscopic investigation called Orientation Microscopy is already well known in scanning electron microscope (SEM). Spatial resolution in SEM causes the limitation for investigation of fine grained and highly deformed materials. Needs for investigation in nanoscale contribute to development of an appropriate method for transmission electron microscope (TEM). Automated acquisition and indexing Kikuchi diffraction pattern, necessary for creation of orientation maps in TEM, cause more difficulties than in SEM. In order to solve the problem, the author developed and tested three methods for automatic Kikuchi lines detection. The first method is based on directional image filtration and scanning the entire image with a specially designed mask. This method yields good results but is relatively slow. The second method make use of the modified Hough transform. The author introduced some modifications to the standard Hough transform in order to achieve better performance in the presence of noise and other artifacts. The third method make also use of the Hough transform but it also uses more advanced methods for peaks detection in the Hough space as well as additional results verification. The third method was implemented in C++ and extensively tested by the author. The result show that this method outperforms the previous two methods.