|Article title||ALGORITHMS FOR ELECTRO-OPTICAL SYSTEMS OF THE COMPLEX FOR PROTECTION THE RESTRICTED AREAS FROM UNMANNED AERIAL VEHICLES|
|Authors||V. A. Bondarenko, G. E. Kaplinskiy, V. A. Pavlova, V. A. Tupikov|
|Section||SECTION I. PROSPECTS FOR APPLICATION OF ROBOT SYSTEMS|
|Month, Year||01, 2018 @en|
|Index UDC||004.896: 004.932.72'1|
|Abstract||This article considers approaches to the development of computer vision algorithms for electro-optical systems. The aim of the study is to expand the scope and improve the effectiveness of electro-optical systems in tasks of protecting the closed areas from unauthorized intrusion of unmanned aerial vehicles (UAV). To solve the problems of automatic detection, recognition and auto-tracking of the UAV, a combined optoelectronic system including a wide-field (panoramic) and a narrow-field subsystem of optical-electronic monitoring was developed. The proposed system is a part of specialized complex of means for protection of closed areas, including the means of radio-electronic monitoring, radio-electronic counteraction and operator’s automated workspace. A step-by-step solution of the main image processing and object recognition tasks is considered in detail. From the framework of algorithms of the wide-field subsystem of electro-optical monitoring, an algorithm for reducing optical distortions from the use of wide-angle lenses and an algorithm for automatic object detection in a panoramic image are considered. The task of automatic object detection is divided into two stages. At the first stage, the difference between the current panoramic image and the panoramic image of the background situation is determined, which is formed by the accumulation of panoramic images. At the second stage of the detection algorithm, the author"s algorithm for auto-tracking is used to filter and track suspicious zones, determined at the previous stage. This combination of algorithms makes it possible to detect and perform a primary selection of suspicious objects whose angular coordinates are transmitted to a narrow-field electro-optical monitoring subsystem for further analysis. Within the framework of algorithms of a narrow-field subsystem of electro-optical monitoring, the solution of problems of automatic detection, recognition and auto-tracking of objects is considered. The algorithm of multispectral fusion of television and thermal imaging images is described. To solve the problem of automatic recognition of the UAV, it was suggested to use a convolutional artificial neural network. The selected architecture of the artificial neural network showed high classification accuracy and speed of operation. The developed algorithmic framework was tested in the electro-optical systems of the complex of protection of objects from unmanned aerial vehicles and demonstrated high efficiency.|
|Keywords||Electro-optical systems; algorithm; image processing; automatic object detection; auto-tracking; object recognition; artificial neural network.|
|References||1. Myasnikov E.V. Ugroza terrorizma s ispol'zovaniem bespilotnykh letatel'nykh apparatov: tekhnicheskie aspekty problemy [Terrorist threat using unmanned aerial vehicles: technical aspects of problem analysis], Tsentr po izucheniyu problem razoruzheniya, energetiki i ekologii pri MFTI. Dolgoprudnyy [Center of disarming, energetic and ecology problems research at MPTU. Dolgoprudniy], 2004, 29 p.
2. Luis Alvarez, Luis Gomez, and J. Rafael Sendra. Algebraic Lens Distortion Model Estimation, Image Processing on Line, 2010, No. 11, pp. 1-10.
3. Tsai R.Y. A versatile camera calibration technique for high-accuracy 3D machine vision metrology using off-the-shelf TV cameras and lenses, IEEE Journal of Robotics and Automation, RA-3(4), 1987, pp. 323-344.
4. Gonzalez R.C., Woods R.E. Digital Image Processing. 2nd ed. Prentice-Hall, 2002, pp. 91-94.
5. Furman Ya.A., Yur'ev A.N., Yanshin V.V. Tsifrovye metody obrabotki i raspoznavaniya binarnykh izobrazheniy [Means of digital binary image processing and recognition]. Krasnoyarsk: Izd-vo Krasnoyarskogo universiteta, 1992, 248 p.
6. Tupikov V.A., Pavlova V.A., Bondarenko V.A., Aleksandrov V.A. Sposob avtomaticheskogo obnaruzheniya ob"ektov na morskoy poverkhnosti v vidimom diapazone [Algorithm of automatic helicopter type drone landing using onboard optical-electronic system], Izvestiya TulGU. Tekhnicheskie nauki [Izvestiya TulSU. Engineering Sciences], 2016, No. 11-3, pp. 105-121.
7. Sauvola J., Pietikainen M. Adaptive document image binarization, Pattern Recognition, 2000, No. 33, pp. 225-236.
8. Bradley D., Roth G. Adaptive Threshholding Using Integral Image, Journal of Graphics Tools, 2007, Vol. 12, Issue 2, pp. 13-21.
9. Shafait F., Keysers D., Breuel T.M. Efficient implementation of local adaptive thresholding techniques using integral images, Proc. SPIE 6815, Document Recognition and Retrieval XV, 681510, 2008. Doi: 10.1117/12.767755; https://doi.org/10.1117/12.767755.
10. LeCun Y., Boser B., Denker J.S., Henderson D., Howard R.E., Hubbard W. and Jackel L.D. Backpropagation Applied to Handwritten Zip Code Recognition, Neural Computation, 1989, No. 1 (4), pp. 541-551.
11. Goodfellow I., Bengio Y., Courville A. Deep Learning. MIT Press, 2016, 781 p.
12. Deng L. and Yu. D. Deep Learning: Methods and Applications, Foundations and Trends in Signal Processing, 2013, Vol. 7, nos. 3–4, pp. 197-387.
13. Ioffe, Sergey, and Christian Szegedy. Batch normalization: Accelerating deep network training by reducing internal covariate shift, arXiv preprint, 2015. arXiv:1502.03167.
14. Szegedy C. et al. Going deeper with convolutions, Proceedings of the IEEE conference on computer vision and pattern recognition, 2015, pp. 1-9.
15. Kingma D. P., Ba J. Adam: A method for stochastic optimization, ICLR, 2015, pp. 1-15.
16. Frolov V.N., Tupikov V.A., Pavlova V.A., Aleksandrov V.A. Metody informatsionnogo sovmeshcheniya izobrazheniy v mnogokanal'nykh optiko-elektronnykh sistemakh [Informational image fusion methods in multichannel optoelectronic systems], Izvestiya TulGU. Tekhnicheskie nauki [Izvestiya TulSU. Engineering Sciences], 2016, No. 11-3, pp. 95-104.
17. Teterin V.V. i dr. Metod kompleksirovaniya informatsii ot mnogokanal'noy sistemy s ispol'zovaniem veyvlet-spektrov [Method for fusing the information from multichannel system using wavelet specters], Opticheskiy zhurnal [Optical journal], 2006, Vol. 73, No. 10, pp. 47-52.
18. Petukhov A.P. Vvedenie v teoriyu bazisov vspleskov [Introduction to basis splash theory]. Saint Petersburg: Izd-vo SPbGTU, 1999, 132 p.
19. Vorob'ev V.I., Gribunin V.G. Teoriya i praktika veyvlet-preobrazovaniya [Theory and practice of wavelet transform], VUS [CMU], 1999, pp. 1-204.
20. Baklitskiy V.K. Korrelyatsionno-ekstremal'nye metody navigatsii i navedeniya [Correlation-extreme means of navigation and aiming]. Tver': TO «Knizhnyy klub», 2009, 360 p.
21. Alpatov B.A., Babayan P.V., Balashov O.E., Stepashkin A.I. Metody avtomaticheskogo obnaruzheniya i soprovozhdeniya ob"ektov [Methods of automatic object detection and tracking], Obrabotka izobrazheniy i upravlenie [Image processing and control]. Moscow: Radiotekhnika, 2008, 176 p.