UNIV. NEWS. NORTH-CAUCAS. REG. TECHNICAL SCIENCES SERIES. 2022; 3: 48-56
http://dx.doi.org/10.17213/1560-3644-2022-3-48-56
STRUCTURAL APPROXIMATION ALGORITHMS OF PRELIMINARY METROLOGICAL CONTROL OF CELLULAR SURFACES OF INDUSTRIAL PRODUCTS
Manilo Maksim K. – Graduate Student, Department «Computer Engineering Software», maksim.manilo.95@mail.ru
Sinetsky Roman M. – Candidate of Technical Sciences, Associate Professor, Department «Computer Engineering Software», rmsin@srspu.ru
Abstract
The article is devoted to the application of the structural approximation method for the recognition of two-dimensional images of cellular surfaces of industrial products in the problem of preliminary metrological photo control. High requirements are imposed on the quality of manufacture of products, consisting in the accuracy of observing the geometric dimensions and the location of elements, since the strength of the product significantly depends on this. As a rule, manufacturing quality control is carried out using laser scanning, which has high accuracy, but is very time-consuming. Preliminary photo-graphic control, having a lower accuracy, allows at an early stage without significant time costs to reject products with obvious defects, without sending them for thorough scanning. The paper presents the basic concepts of the used structural theory of images, the structural-approximation approach, describes three models of images: deformed, describing the observed photographic image of the product; ideal, perceived as the standard of the image; approximation, which, according to the given criteria, is close to the ideal and deformed images. Algorithms for the formation of a deformed image, synthesis and analysis of the approximation image are given.
Keywords: configurations, pattern recognition, structural approximation, metrological control, raster image analysis, image processing
Full text: [in elibrary.ru]
References
- Manilo M.K. Structural-approximation approach in the problem of photo-control of cellular surfaces of industrial products. Actual scientific research in the modern world: proceedings of the LXXX International Scientific Conference, December 2021, Pereyaslav-Khmelnitsky. Journal - Pereyaslav-Khmelnitsky. 2021;12(80):91-96. (In Russ.).
- Sinetsky R.M., Manilo M.K. Application of the method of structural approximation in the problem of photocontrol of cellular surfaces of industrial products. Results of research - 2020. Proceedings of the V National Conference of the faculty and scientists of the SRSPU (NPI). Novocherkassk, 2020. Pp. 37-40.
- Grenander U. Lectures on the theory of images. Synthesis of images. Book. I. Moscow: Mir; 1979.
- Manilo M.K. Methods of structural approximation in various fields of application. International scientific and technical conference of young scientists of BSTU named after. V.G. Shukhova. Proceedings of the conference. Belgorod. 2021. Pp. 3673-3677.
- Gavrikov M.M., Sinetsky R.M. Algorithm for one-dimensional structural-approximation analysis of images of cellular surfaces in the problem of input metrological control of industrial products. Inrzenernyi vestnik Dona = Engineering Bulletin. Don (electronic scientific journal). 2020; (6). (In Russ.).
- Gavrikov M.M. Structural approximation and recognition of one-dimensional temporal patterns. Concept and applications. Izvestiya Vysshihkh Uchebnykh Zavedenii. Elektromekhanika = Russian Electromechanics. 2003; (6): 52-60. (In Russ.).
- Bellman R., Dreyfus S. Applied problems of dynamic programming.Moscow: Nauka; 1965. 460 p.
- Grenander U. Lectures on the theory of images. Image analysis. Book. II. Moscow: Mir; 1981.
- Gavrikov M.M., Sinetsky R.M., Knyazev D.N. Technology of metrological control of cellular cylindrical surfaces of large-sized products. Izvestiya Vysshihkh Uchebnykh Zavedenii. Elektromekhanika = Russian Electromechanics. 2016; (4):100-105. DOI: 10.17213/0136-3360-2016-4-100-105. (In Russ.).
- Fukunaga K. Introduction to Statistical Pattern Recognition. 2nd edition. Academic Press; 1990. 592 p.