Published date: Oct 08 2018

Journal Title: KnE Engineering

Issue title: Breakthrough directions of scientific research at MEPhI: Development prospects within the Strategic Academic Units

Pages: 305–310

DOI: 10.18502/keg.v3i6.3008

N S Kukinn.s.kukin@mail.ru

I O Korchevaya

The article presents solution to the problem of objects detection by series of their images without any prior information about outer characteristics (shape, color, etc.). The presented approach is based on detecting connections between separate points of objects. The authors developed an algorithm that uses movements of a fixed number of points measured at discrete instants of time as input data. It was tested using real experimental data. As a result, a matrix of links was obtained, which can be used for identifying separate objects. In future articles, the authors will offer the improvement of the proposed approach for quantifying the nature of connections between different points of a single object. It will help determine the object structure, it’s weak and strong parts, internal defects, etc.

Keywords: series of images, points movements, links between points

[1] Glumov, N. I., Myasnikov, E. V., Kopenkov, V. N., et al. (2008). Metod bystroj korrelyacii s ispolzovaniem ternarnyx shablonov pri raspoznavanii obektov na izobrazheniyax [The method of rapid correlation using ternary patterns when recognizing objects in images]. Kompyuternaya Optika, vol. 32, no. 3, pp. 277–282.


[2] Zavyalov, R. A. and Senyushkin, N. S. (2012). Avtomaticheskoe raspoznavanie obrazov bortovymi opticheskimi sistemami mnogocelevyx kompleksov monitoringa [Automatic recognition of images by onboard optical systems of multi-purpose monitoring complexes]. Vestnik UGATU, vol. 16, no. 6(51), pp. 12–16.


[3] Knyazev, B. A. and Cherkasskij, V. S. (2008). Diskretnoe preobrazovanie Fure – kak eto delaetsya [Discrete Fourier Transform – How It’s Done]. Vestnik NGU, Fizikavol, vol. 3, no. 4, pp. 74–86. Seriya.


[4] Mesteckij, L. M. (2002–2004). Matematicheskie metody raspoznavaniya obrazov [Mathematical methods of pattern recognition]. Kurs lekcij, MGU, VMiK, kafedra «Matematicheskie metody prognozirovaniya».


[5] Muradova, A. R. and Kukin, N. S. (2016). Detektirovanie peremeshcheniya ob’ekta na rasstoyanie men’she odnogo pikselya po serii ego izobrazhenij [Detection of the movement of an object at a distance of less than one pixel in a row of its images]. Izvestiya Instituta inzhenernoj fiziki, vol. 4, no. 42, S. 36–39.


[6] Nikitin, A. K. and Kukin, N. S. (2016). Izvlechenie unikalnoj informacii iz razlichnyx izobrazhenij obekta v odnix i tex zhe usloviyax [Extraction of unique information from various images of the object under the same conditions]. Izvestiya Instituta inzhenernoj fiziki, vol. 4, no. 42, pp. 32–35.


[7] Shemagina, O. V. (2016). Metody raspoznavaniya obrazov v zadache analiza izobrazhenij i video po soderzhaniyu [Methods of pattern recognition in the task of analyzing images and video by content]. Avtoref. dis. kand. texn. nauk. - Nizhnij Novgorod., p. 28.