Automatic recognition of the structures of ferritic-pearlitic steels on the images, obtained by the method of atomic-force microscopy

The problem of automatic recognition of the structures of samples images of ferritic-pearlitic steels, obtained by the method of atomic force microscopy, is considered. For its solution the algorithm, which includes image pre-processing, segmentation, calculation of the image parameters and the definition according to the received parameters of the most probable state of the tested steel sample by means of the supporting vectors method, is proposed. The possibility of accuracy increasing of classification by means of determining of the optimal set of the images parameters is considered. The algorithm, described in this article, is realized in the form of software module. The examples of the experimental data processing are given.