Characterization of objects arrangement in image analysis of physical systems

Abstract In some fields of physics the systems consisting of individual objects are studied. One of the most important tasks of image analysis of such systems is the quantitative characterization of the degree of objects arrangement. For the description of spatial distribution of objects several morphological methods are typically used. However, the interpretation of derived characteristics is often difficult. In order to contribute to proper interpretation of standard morphological characteristics and to test less known algorithms for the quantitative description of the objects arrangement, a computer experiment was prepared. Several types of simulated structures corresponding to various types of experimental data in physics were generated and analysed by six algorithms of mathematical morphology and Fourier optics. The results of sensitivity analysis of these algorithms were applied to the transmission electron microscope images of thin metal films.