Abstract and keywords
Abstract (English):
We consider an approach to solving the problem of image analysis of complex formalized structures of materials of critical parts. The concept of the variant structure of the material is defined. Methods of managed image segmentation, building and training network models are developed. The result of evaluating the coating structure by plasma spraying is shown.

Keywords:
variant structure, automation of material structure analysis, image processing, network model, training on precedents
References

1. Emel'yanov V.A. Modelirovanie neyronnyh setey raspoznavaniya metallograficheskih izobrazheniy dlya diagnostiki sostoyaniya staley // Elektrotehnicheskie i komp'yuternye sistemy. 2013. № 12. S. 125 - 131.

2. Aristov G.V., Klyuev A.V. Raspoznavanie i klassifikaciya mikrostruktury metallov i splavov s ispol'zovaniem glubokih neyronnyh setey / V sbornike: GraphiCon 2017. Trudy 27-y Mezhdunarodnoy konferencii po komp'yuternoy grafike i mashinnomu zreniyu. Permskiy gosudarstvennyy nacional'nyy issledovatel'skiy universitet. 2017. S. 180 - 183.

Login or Create
* Forgot password?