ALGORITHM FOR OBJECTS’ DETECTING AND CLASSIFICATION ON A HETEROGENEOUS BACKGROUND 

Chernikov Andrey Andreevich, Postgraduate student, Novosibirsk State Technical University (20 K. Marksa avenue, Novosibirsk, Russia), E-mail: ancher1994@gmail.com
Purtov Anton Igorevich, Postgraduate student, Novosibirsk State Technical University (20 K. Marksa avenue, Novosibirsk, Russia), E-mail: a.p.93@mail.ru
Prokof'ev Ivan Valer'evich, Postgraduate student, Novosibirsk State Technical University (20 K. Marksa avenue, Novosibirsk, Russia), E-mail: prokofev.ivan.93@mail.ru
Yushchenko Valeriy Pavlovich, Doctor of engineering sciences, professor, sub-department of autonomous information and control systems, Novosibirsk State Technical University (20 K. Marksa avenue, Novosibirsk, Russia), E-mail: yushhenko@corp.nstu.ru

004.932.2

10.21685/2072-3059-2020-4-4 

Background. The object of the research is an optoelectronic detection system for unmanned aerial vehicles and armored vehicles. The subject of the research is the methods of identifying and classifying a moving object against a complex nonuniform background. The aim of the research is to develop an algorithm for the objects’
detecting and classification of an unmanned aerial vehicle and armored vehicles by an optoelectronic system against a non-uniform background in real time.
Materials and methods. The presented studies were carried out using video image processing methods to select an object and neural networks to classify an object.
The algorithm is developed in the Python programming language using the  OPENCV computer vision library.
Results. A method for identifying and classifying an unmanned aerial vehicle and armored vehicles against a complex dynamic background is proposed. The algorithm uses a Harris angle detector to detect objects in the background of images. Created and trained a neural network for fast object classification.
Conclusion. The proposed method can be used to develop an optoelectronic system for detecting a moving unmanned aerial vehicle and armored vehicles against a non-uniform background in real time in the infrared range. Because of the work, it was revealed that the proposed algorithm reliably copes with the detection and classification of a contrasting object located at a distance of up to 2 km from the detection system.

optoelectronic system, unmanned aerial vehicle, armored vehicles, adaptation, neural networks.

1. Li J., Shen Z., Bao L. Automatic targe trecognition IV: Proc. Of SPIE. 1997, vol. 3069, pp. 511–517.
2. Alpatov B. A., Blokhin A. N., Murav'ev V. S. Tsifrovaya obrabotka signalov [Digital signal processing]. 2010, no. 4, pp. 12–17. [In Russian]
3. Borisova I. V., Gorenok V. N., Zakharova V. N. Problemy avtomatizatsii i upravleniya v tekhnicheskikh sistemakh: tr. Mezhdunar. nauch.-tekhn. konf. (g. Penza, 19–22
aprelya 2011 g.) [Automation and control problems in technical systems: proceedings of the International scientific and engineering conference (Penza, April 19-22, 2011)]. Penza: Izd-vo PGU, 2011, vol. 1, pp. 186–190. [In Russian]
4. Ciresan D., Schmidhuber J. IEEE Proceedings of Computer Vision and Pattern Recognition. 2012, pp. 3642–3649.
5. Fedotov N. G., Goldueva D. A., Mokshanina M. A. Izvestiya vysshikh uchebnykh zavedeniy. Povolzhskiy region. Tekhnicheskie nauki [University proceedings. Volga region. Engineering sciences]. 2017, no. 2 (42), pp. 29–41. [In Russian]
6. He K., Zhang X., Ren S., Sun J. Proc. of the IEEE conf. on computer vision and pattern recognition. LasVegas, 2016, pp. 770–778.
7. Sedzhvik R., Ueyn K., Dondero R. Programmirovanie na yazyke Python: uchebnyy kurs: per. s angl. [Python programming: a training course: translated from English].
Saint-Petersburg: OOO «Al'fa-kniga», 2017, 736 p. [In Russian]
8. Gonsales R., Vuds R. Tsifrovaya obrabotka izobrazheniy [Digital imaging]. Moscow: Tekhnosfera, 2005, 1072 p. [In Russian]
9. Otchet o NIR «Sostoyanie i perspektivy razvitiya matrichnykh FPU dlya mnogospektral'nykh GSN», shifr «Neyron», № gos. reg. U90056, inv. № E79158 [Report on
the Scientific Research “State and development prospects of matrix FPUs for multispectral GOS”, cipher “Neuron”, state registration No. U90056, inv. No. E79158]. FGUP
«Sibirskiy nauchno-issledovatel'skiy institut opticheskikh sistem», nauchnyy ruk. Legkiy V. N. Novosibirsk, 2007, 53 p. [In Russian]
10. Gruzman I. S., Kirichuk V. S., Kosykh V. P., Peretyagin G. I., Spektor A. A. Tsifrovaya obrabotka izobrazheniy v informatsionnykh sistemakh [Digital image processing in information systems]. Novosibirsk: Izd-vo NGTU, 2002, 352 p. [In Russian]