Main Article Content
The subject of the study in the article is the neural network method of object recognition on multispectral Earth remote sensing (ERS). The goal providing automatic recognition of objects illegal exploitation of natural resources in multispectral ERS images. The task is formulation of the method of intellectual processing of ERS data, which implements automatic recognition of objects of illegal use of natural resources on multispectral ERS images by using a convolutional neural network.. Analysis of the problems of methods and algorithms for processing multispectral aerospace images has shown that it is most promising to use flexible algorithms that adapt to changing conditions for observing search objects. One of the most promising technologies of the implementation of such algorithms is the use of neural networks. The selection of convolutional neural networks for solving the recognition problem is related to the ability of these networks, under the condition of correct training, to recognize objects under difficult observation conditions and when the observed object. Conclusions: the neural network method of intellectual processing of multispectral images is proposed. The algorithm for constructing this network is considered, the practical scope of the proposed method is chosen and the results of its program implementation are shown. The obtained results made it possible to conclude that the proposed algorithm is working and are the basis for further research into the development and implementation of processing algorithms for multispectral images in ERS systems.
Baklanov, A.I. (2010), Analyz sostoianyia y tendentsyy razvytyia system nabliudenyia visokoho y sverkhvisokoho razreshenyia [Analysis of the state and tendency of development of observation systems of high and ultrahigh resolution], Vestnik Samarskoho Gosudarstvennogo Aerokosmycheskogo Universiteta [Bulletin of the Samara State Aerospace University], Samara, No 2, pp. 80-91.
Potikha, A. (2016), Problema vydobutku burshtynu: suchasnyi stan ta perspektyvy vyrishennia [The problem of amber extrac-tion: the current state and prospects of solution], Ukraina: podii, fakty, komentari [Ukraine: Events, Facts, Comments], Kyiv, No 5, pp. 36–44, available at: http://nbuviap.gov.ua/images/ukraine/2016/ukr5.pdf (last accessed January 23, 2017).
Kashkin, V.V. and Sukhinin, A.I. (2008), Tsifrovaya obrabotka aerokosmicheskikh izobrazheniy [Aerospace images digital processing], IPK SFU, Krasnoyarsk, 121 p.
Gonzalez, R. and Woods, R. (2008), Digital Image Processing, Pearson Prentice Hall, 954 p.
LeCun, Y. and Bengio, Y. (1995), Convolutional networks for images, speech and time series, The handbook of brain theory and neural network, Vol. 3361, No 10, MIT Press, pp. 276–279.
Gorbachevskaya, E.N. and Krasnov, S.S. (2015), Istoryia razvityia neironnikh setei, [History of neural networks development], Vestnik Volzhskogo universiteta imeni V.N. Tatyshcheva [Bulletin of the Volga University named after V.N. Tatischev], VUiT, Tolyatti, No 1 (23), pp. 52-56.
LeCun, Y., Kavukcuoglu, K.. and Farabet, C. (2010), Convolutional Networks and Applications in Vision, Proceedings of 2010 IEEE International Symposium on Circuitsс and Systems (ISCAS’10), IEEE, Paris, pp. 253–256.
Galushkin, A.I. (2012), Neironnye seti. Osnovy teorii, [Neural networks. Fundamentals of the theory], Horiachaia Lynyia – Telekom, Moscow, 496 p.
Barannyk, V.V. and Podorozhniak, A.A. (2014), Metod intellektualyzatsyy obrabotky dannikh v bortovoi apparature systemi dystantsyonnoho zondyrovanyia zemly [The method of data processing intellectualization in on-board equipment of the Earth remote sensing system], Suchasna spetsialna tekhnika [Modern special technique], DNDI MVS Ukrainy, Kyiv, No 2 (37), pp. 5-13.
Podorozhniak, A.O. Pribyliev, Yu.B. and Torokhtii, D.I. (2014), Metod intelektualnoi obrobky danykh dystantsiinoho zonduvannia Zemli [The method of intellectual data processing of remote sensing], Systemy obrobky informatsii [Information processing systems], KhUPS, Kharkiv, No 2 (118), pp. 48-51.
Podorozhniak, A.O., Liubchenko, N.Yu. and Lagoda, O.D. (2015), Metod intelektualnoi obrobky multyspektralnykh zobrazhen [The method of intellectual multispectral image processing], Sistemi obrobki informatsii [Information processing systems], KhUPS, Kharkiv, No 10 (135), pp. 123-125.
Podorozhnyak, A.A. (2014), Metod viyavlenyia obektov interesa pry obrabotke dannikh v systeme dystantsyonnoho zondyrovanyia zemly, [Method of interest objects detection while processing data in the system of earth remote sensing], Informacijno-kerujuchi systemy na zaliznychnomu transporti [Information management systems in the railway transport], UkrDAZT, Kharkiv, No 4, pp. 60-64.
Digital map service Here.com, available at: https://here.com/en (last accessed January 23, 2017).
Bondarchuk, V.K., Podorozhniak, A.O. and Liubchenko, N.Yu. (2016), Vykorystannia zghortkovykh neiromerezh dlia rozpiznavannia obiektiv na zobrazhenniakh [Use of convolutional neural networks to recognize objects in images], Tezy 16 mizhnarodnoi naukovo-tekhnichnoi konferentsii Problemi informatiky i modeliuvannia [Proceedings of 16th International scientific and technical conference Problems of Informatics and Modeling], NTU "KhPI", Kharkiv, p. 54.