Multilevel analysis of the relief of a surface sample obtained by atomic force microscopy techniques

Atomic force microscopy is used to study the structure of a material and its local mechanical properties on the nanoscale. This information can be applied to analyze the macroscopic homogeneity nanocomposites (quality of mixing with filler), the mechanical properties of which depend directly on the distribution of filler particles throughout the material. Hence, it is necessary to identify filler particle aggregates visible on the surface of a sample. For this purpose, a technique capable of identifying objects of a specific size on the curved surface is proposed. The method is based on the decomposition of the initial topographic image into the sum of several reliefs using a linear averaging spatial filter. Each relief is identified so that it contains certain-size objects and can be analyzed separately. It is suggested that this procedure be called multilevel visualization or multilevel analysis. This paper is focused on the decomposition of the initial topographic image into three components: relief with objects of high curvature surface, low curvature relief and relief with objects of medium curvature surface. The technique is intended to study nanocomposites created using granular filler. The efficiency of the developed method has been verified by testing a specially designed surface, which is much like as the real surface of the material analyzed by AFM. Parameters affecting the accuracy of object identification are determined. Recommendations for practical application of the multilevel visualization to the analysis of experimental data are given.