Computer Optics

863-870

Authors: V.V. Kotlyar, S.S. Stafeev, A.G. Nalimov, A.A. Kovalev, A.P. Porfirev

Using two identical microobjectives with a numerical aperture NA = 0.95, we experimentally demonstrate that the on-axis intensity near the tight focal spot of an optical vortex with a topological charge 2 is zero for right-handed circular polarization and nonzero for left-handed circular polarization. This serves to confirm that in the latter case there is a reverse energy flow on the optical axis, as testified by a very weak local maximum (the Arago spot) detected at the center of the measured energy flow distribution, caused by diffraction of the direct energy flow by a 300 nm circle (the diameter of a reverse energy flow tube). The comparison of numerical and experimental intensity distributions shows that it is possible to determine the diameter of the reverse energy flow "tube", which is equal to the distance between the adjacent intensity nulls. For NA = 0.95 and a 532 nm incident wavelength, the diameter of the on-axis reverse energy flow "tube" is measured to be 300 nm. It is also experimentally shown that when an optical beam with second-order cylindrical polarization is focused with a lens with NA = 0.95, there is a circularly symmetric energy flow in the focus with a very weak maximum in the center (the Arago spot), whose distribution is determined by diffraction of the direct energy flow by a 300 nm circular region, where the energy flow is reverse. This also confirms that in this case, there is a reverse energy flow on the optical axis.

876-882

Authors: C.N. Alexeyev, E.V. Barshak, B.P. Lapin, M.A. Yavorsky

In this paper, we studied an effect of strong evanescent coupling on the structure of normal modes in a system of parallel ideal multimode optical fibers. Using the formalism of the degenerate perturbation theory and a scalar waveguide equation for this system, analytical expressions of higher-order supermodes and their propagation constants have been determined. We have shown that the structure of modes in the case of strong evanescent coupling coincides with the structure of normal modes for weakly coupled parallel fibers. We have demonstrated that in the presence of strong coupling, expressions for corrections to the scalar propagation constant are modified, deducing them analytically.

893-900

Authors: A.V. Dyshlyuk, A.A. Bogdanov, O.B. Vitirk

In this work, we demonstrate a simple analytical approach to the problem of surface plasmon polaritons excitation with a metallic nanoantenna placed above a metal surface. The method uses the reciprocity theorem and is similar to the calculation of amplitudes of dielectric waveguide modes excited by a current distribution. To maximize clarity of the demonstration, we formulate the problem in a simple two-dimensional geometry. The analytical results are shown to agree well with the numerical solution obtained by finite elements in frequency domain and finite difference in time domain methods.

909-916

Authors: V.D. Zaitsev, S.S. Stafeev

Using the finite difference method implemented in the COMSOL Multiphysics software package, the focusing of laser radiation by dielectric prisms with a triangular profile was numerically investigated. It was shown that two-dimensional triangular prisms make it possible to focus light in free space into spots with dimensions smaller than the scalar diffraction limit. In particular, a silica glass prism with a base width of 60 μm and a height of 28.5 μm forms a photonic nanojet with a maximum intensity of 6 times the intensity of the incident radiation and a width of FWHM=0.38λ. A prism from barium titanate with a base width of 60 μm and a height of 20 μm allows to obtain a photonic nanojet with the same width (0.38λ) and a maximum intensity 5 times the intensity of the incident radiation. The size of the focal spot can be reduced further if the height of the prism is selected so that the maximum intensity is located inside the material of the prism. For example, a barium titanate prism with a height of 21 μm and a base width of 60 μm forms a focal spot with a width of FWHM=0.25λ.

923-930

Authors: I.A. Rodin, S.N. Khonina, P.G. Serafimovich, S.B. Popov

In this work, we carried out training and recognition of the types of aberrations corresponding to single Zernike functions, based on the intensity pattern of the point spread function (PSF) using convolutional neural networks. PSF intensity patterns in the focal plane were modeled using a fast Fourier transform algorithm. When training a neural network, the learning coefficient and the number of epochs for a dataset of a given size were selected empirically. The average prediction errors of the neural network for each type of aberration were obtained for a set of 15 Zernike functions from a data set of 15 thousand PSF pictures. As a result of training, for most types of aberrations, averaged absolute errors were obtained in the range of 0.012 – 0.015. However, determining the aberration coefficient (magnitude) requires additional research and data, for example, calculating the PSF in the extrafocal plane.

937-943

Authors: S.M. Borzov, O.I. Potaturkin

Classification of the land cover types from multi- and hyperspectral (HS) imagery is traditionally carried out on the basis of analysis of scatter plots of pixel values in a multidimensional feature space, which are used as brightness in individual channels. To increase the reliability of HS image classification, approaches are used based on simultaneously accounting for the characteristics of each pixel and the nearest-neighbor pixels, i.e., on the joint analysis of spectral and spatial features. The pixel neighborhood analysis is performed at various stages of the classification process. In this work, using a test hyperspectral image, the efficiency of spectral-spatial data classification methods that take into account spatial information at various stages of processing is studied. Special attention is paid to selecting the size of the spatial processing core. It is shown that the best results are obtained by combining pre-processing of raw data before performing the procedures of pixel-by-pixel spectral classification and post-processing of the resulting maps. Prospects of multi-scale smoothing of initial images, with the increase of the number of spectral-spatial features being multiple of the number of the scales, are shown.

951-958

Authors: O.S. Seredin, A.V. Kopylov, E.E. Surkov

Accurate and reliable real-time fall detection is a key aspect of any intelligent elderly people care system. A lot of modern RGB-D cameras can provide a skeleton description of a human figure as a compact pose presentation. This makes it possible to use this description for further analysis without access to real video and, thus, to increase the privacy of the whole system. The skeleton description reduction based on the anthropometrical characteristics of a human body is proposed. The experimental study on the TST Fall Detection dataset v2 by the Leave-One-Person-Out method shows that the proposed skeleton description reduction technique provides better recognition quality and increases the overall performance of a Fall-Detection System.

968-977

Authors: M.O. Kalinina, P.L. Nikolaev

Nowadays deep neural networks play a significant part in various fields of human activity. Especially they benefit spheres dealing with large amounts of data and lengthy operations on obtaining and processing information from the visual environment. This article deals with the development of a convolutional neural network based on the YOLO architecture, intended for real-time book recognition. The creation of an original data set and the training of the deep neural network are described. The structure of the neural network obtained is presented and the most frequently used metrics for estimating the quality of the network performance are considered. A brief review of the existing types of neural network architectures is also made. YOLO architecture possesses a number of advantages that allow it to successfully compete with other models and make it the most suitable variant for creating an object detection network since it enables some of the common disadvantages of such networks to be significantly mitigated (such as recognition of similarly looking, same-color book coves or slanted books). The results obtained in the course of training the deep neural network allow us to use it as a basis for the development of the software for book spine recognition.

985-989

Authors: Z.M. Cai

The division of network community is an important part of network research. Based on the clustering algorithm, this study analyzed the partition method of network community. Firstly, the classic Louvain clustering algorithm was introduced, and then it was improved based on the node similarity to get better partition results. Finally, experiments were carried out on the random network and the real network. The results showed that the improved clustering algorithm was faster than GN and KL algorithms, the community had larger modularity, and the purity was closer to 1. The experimental results show the effectiveness of the proposed method and make some contributions to the reliable community division.

685-690

Authors: S.S. Stafeev, V.V. Kotlyar

In this paper, we numerically investigated tight focusing of cylindrical vector beams of the sec-ond order using Richards-Wolf formulae. It was shown that intensity rings where the Poynting vector was equal to zero appeared not only in the focal plane but also in nearby planes. For example, a lens with numerical aperture NA=0.95 was shown to generate periodical toroidal vortices with a 0.8-mkm period along the z-axis at a distance of about 0.45 mkm from the axis. The vortices were generated pairwise, with the closest-to-focus vortex having clockwise helicity and the subsequent being anti-clockwise. The vortices were also characterized by saddle points. When focusing an optical beam passed through a narrow annular aperture, no toroidal vortices were observed.

699-706

Authors: A.G. Nalimov, E.S. Kozlova

It has been shown theoretically and numerically that in the sharp focus of a circularly polarized optical vortex, the longitudinal component of the spin angular momentum vector is inverted. Moreover, if the input light to the optical system is left-hand circularly polarized, it has been shown to be right-hand polarized in the focus near the optical axis. Since this effect occurs near the focus where a backward energy flow takes place, such an inversion of the spin angular momentum can be used to detect the backward energy flow.

712-720

Authors: E.V. Byzov, S.V. Kravchenko, M.A. Moiseev, L.L. Doskolovich

A method for designing an optical element with two free-form surfaces generating a prescribed illuminance distribution in the case of an extended light source is considered. The method is based on the representation of the optical element surfaces by bicubic splines and on the subsequent optimization of their parameters using a quasi-Newton method implemented in the Matlab software. To calculate the merit function, a version of the ray tracing method is proposed. Using the proposed method, an optical element with record characteristics was designed: the ratio of the element height to the source size is 1.6; luminous efficiency is 89.1 %; uniformity of the generated distribution (the ratio of the minimum and average illuminance) in a given square region is 0.92.

728-736

Authors: A.V. Pavlov

A 6f-scheme of Fourier holography with resonant architecture is considered, which implements memory replenishment with new information that contradicts the previously recorded. It is shown that the low-frequency attenuation due to the nonlinearity of the exposure characteristics of holographic recording media in the initial reference holographic image recorded in a narrow filtering range corresponding to the degradation in the correlation plane of the global maximum of the autocorrelation function below the lateral maxima leads to the response instability – an intermittent mode. It is shown that the intermittent mode corresponds to the restructuring of the autocorrelation function of a composite standard recorded in holograms from one range of values of the approximation model parameters to another. It is shown that the correlation length of the composite image recorded in holograms is an order parameter of the system; its rapid change precedes the loss of response stability and the transition to an unstable regime with intermittency. The results of numerical simulation are presented.

746-756

Authors: B.A. Alpatov, P.V. Babayan, M.D. Ershov

The paper discusses different approaches to image and video processing aiming to solve the problems of detecting, tracking and estimating the parameters of moving objects. The developed algorithms for solving these problems are described in relation to the field of transport analytics. When developing the algorithms, attention was given to solving the problems on an embedded platform of video surveillance cameras, which imposes restrictions on the computational complexity. The first (basic) algorithm for moving object detection and parameter estimation is based on processing two associated areas of an image. This algorithm includes a computationally efficient adaptive procedure for evaluating and updating the background component of an image. The procedure is based on the physics of the process of movement of the object of interest through a processing zone. The second algorithm performs object tracking based on an optical flow method initialized by feature points. The third algorithm is based on object segment tracking and is computationally efficient for the implementation on an embedded platform of intelligent cameras. Results of experimental studies of the proposed algorithms are presented, as well as a comparison with some well-known algorithms. It is shown that tracking algorithms can improve the accuracy of moving object parameter estimation. Tracking also reduces the number of classification errors compared to the basic approach to object detection and parameter estimation.

763-771

Authors: A.V. Kuznetsov, M.V. Gashnikov

We investigate image retouching algorithms for generating forgery Earth remote sensing data. We provide an overview of existing neural network solutions in the field of generation and inpainting of remote sensing images. To retouch Earth remote sensing data, we use image-inpainting algorithms based on convolutional neural networks and generative-adversarial neural networks. We pay special attention to a generative neural network with a separate contour prediction block that includes two series-connected generative-adversarial subnets. The first subnet inpaints contours of the image within the retouched area. The second subnet uses the inpainted contours to generate the resulting retouch area. As a basis for comparison, we use exemplar-based algorithms of image inpainting. We carry out computational experiments to study the effectiveness of these algorithms when retouching natural data of remote sensing of various types. We perform a comparative analysis of the quality of the algorithms considered, depending on the type, shape and size of the retouched objects and areas. We give qualitative and quantitative characteristics of the efficiency of the studied image inpainting algorithms when retouching Earth remote sensing data. We experimentally prove the advantage of generative-competitive neural networks in the construction of forgery remote sensing data.

782-790

Authors: A.V. Vorobev, V.A. Pilipenko, T.A. Enikeev, G.R. Vorobeva

The paper is concerned with an approach to developing a specialized web-GIS based on a microservice architecture that provides analytical control of the disturbed component of geomagnetic field variations, according to observation data from magnetic observatories and variational stations published on the SuperMAG portal . A method of spatial interpolation of geomagnetic data implemented in the proposed web-GIS, together with the proposed scheme for ranking and interpreting them, as well as a visualization method in the form of isolines, allows a user to track the structure, observe the dynamics, identify probable regions, duration and time intervals of the occurrence of extreme geomagnetic disturbances.

797-808

Authors: O.V. Mandrikova, A.A. Stepanenko

A method for calculating the geomagnetic activity index Dst (Dst-index) based on a wavelet model of geomagnetic field variations is proposed. The method allows values of the Dst-index to be automatically obtained with a 1-minute resolution. The method is tested using data from equatorial stations [1]. The paper describes a calculation algorithm and presents estimation results. The calculation results are compared with the classical approach and the Kyoto method [2]. It is shown that the proposed method allows values of the Dst index to be obtained in the on-line mode with an admissible error.

821-829

Authors: E.A. Kochegurova, E.Yu. Repina, O.B. Tsekhan

In this work, a hybrid-forecasting model is proposed. The model includes a recursive penalty P-spline with parameters adaptation based on evolutionary optimization algorithms. In short-term forecasting, especially in real-time systems, the urgent task is to increase the forecast speed without compromising its quality. High forecasting speed has been achieved by an economical computational scheme of a recurrent P-spline with a shallow depth of prehistory. When combined with the adaptation of some parameters of the P-spline, such an approach allows you to control the forecast accuracy.

843-847

Authors: Z.K. Hou

In the development of modern logistics, the role of automated cargo warehousing is gradually reflected, which is essential for the automatic distribution of goods. This paper briefly introduced the automatic location allocation model and the particle swarm optimization (PSO) algorithm used to optimize the model. At the same time, it introduced the concept of genetic operator and multi-group co-evolution to improve the algorithm, and then the simulation analysis of standard PSO and improved PSO was performed on MATLAB software. The results showed that the improved PSO iterated fewer times and get better solution sets; compared with the manual allocation scheme, the improved PSO calculation reduced more warehousing time, lowered more center of gravity height, and improved shelf stability. In summary, the improved PSO algorithm can effectively optimize the automated goods dynamic allocation and warehousing model.

852-856

Authors: B.A. Knyazev, V.S. Pavelyev

The feasibility of generating surface plasmon polaritons carrying orbital angular momentum ("vortex plasmons") on cylindrical conductors by an end-fire coupling technique in the spectral range from 8.5 to 141 μm (~ 2-40 THz) is considered. The front face of the cylinder is illuminated by Bessel beams formed using binary spiral phase axicons, or annual vortex beams formed in the focal plane of an additional lens. Graphs are constructed that reveal the relationship between the waveguide parameters (conductor diameter, which is equal to the diameter of the illuminating beam, and the “twist” angle of the plasmon) and the axicon parameters (the ratio of the axicon period to the radiation wavelength) for the above wavelengths and topological charges of the beams ranging from 1 to 9. The results obtained indicate the possibility of conducting experiments in the long-wavelength range for modeling a plasmon multiplex communication channel.

493-500

Authors: V.V. Kotlyar, A.A. Kovalev, A.P. Porfirev

In contrast to the orbital angular momentum (OAM), which is conserved on free space propagation, the topological charge (TC) of a paraxial optical vortex (OV) is not conserved in the general case. Here, we investigate a Gaussian beam with a fractional TC in the original plane and demonstrate both theoretically and numerically how the TC changes in the course of propagation. Depending on the proximity of the topological charge to an even or odd integer number, an optical vortex with the original fractional TC is shown to behave in a number of different ways. For simple OVs (Laguerre-Gaussian or Bessel-Gaussian modes), TC is conserved both in propagation and after weak phase distortions. An experiment shows that when scattered by a random phase screen, the integer TC of an OV is conserved right up to a random phase variation of π. Therefore, in the case of weak turbulences, it is expedient to measure a discretely varying TC instead of a continuously varying OAM.

510-518

Authors: V.V. Kotlyar, A.A. Kovalev

Here we theoretically obtain values of the topological charge (TC) for vortex laser beams devoid of radial symmetry: asymmetric Laguerre-Gaussian (LG) beams, Bessel-Gaussian (BG) beams, Kummer beams, and vortex Hermite-Gaussian (HG) beams. All these beams consist of conventional modes, namely, LG, BG, or HG modes, respectively. However, all these modes have the same TC equal to that of a single constituent mode n. Orbital angular momenta (OAM) of all these beams, normalized to the beam power, are different and changing differently with varying beam asymmetry. However, for arbitrary beam asymmetry, TC remains unchanged and equals n. Superposition of just two HG modes with the adjacent numbers (n, n+1) and with the phase retardation of (pi)/2 yields a modal beam with the TC equal to – (2n+1). Numerical simulation confirms the theoretical predictions.

533-539

Authors: V.A. Burdin, A.V. Bourdine, O.Yu. Gubareva

This paper presents the results of an analysis of the necessary propagation conditions in a step-index optical fiber with a Kerr nonlinearity of two modes, LP01 and LP11, during the transmission of high-power optical pulses. All results were obtained by solving a system of two nonlinear equations for these modes, obtained by the Gauss approximation method, and the subsequent use of a procedure for refining estimates using the mixed finite elements method. The necessary conditions are determined, estimates of the boundaries for the range of normalised frequencies for which they are fulfilled are obtained, and an approximate formula is proposed for estimating the upper limit of this range.

547-551

Authors: S.A. Shoydin

A method for the synthesis of holograms that reconstruct a 3D image using two 2D images, including a topographic map of the surface of a 3D object and photographs of the texture of its surface obtained using coherent and incoherent radiation is proposed. The method is suitable for transmitting information about a 3D object via a communication channel at a distance and subsequent synthesis of holograms at a distance from the holography object. It can be used in augmented reality devices, when creating holographic cameras and 3D video cameras, in 3D projectors and microscopes, devices with a synthesized aperture. The proposed method for the remote hologram synthesis is suitable for reproducing 3D holographic images of distant objects at different parts of the electromagnetic spectrum. Examples of 3D image restoration in the infrared spectral range are given. The method is also suitable for the synthesis of holograms in the UV, THz, and even radio range.

561-567

Authors: A.G. Nalimov, S.S. Stafeev

A force and a torque exerted on an elliptical dielectric particle in the focus of a spherical circularly polarized laser beam are considered. The numerical simulation is conducted using a diffraction field obtained by an FDTD method, with the force and torque derived using a Maxwell’s stress tensor. It is shown that an optical torque is exerted on the center of an elliptical particle put in the focus of a circularly polarized spherical wave, making it rotate around the optical axis. The rotation occurs when the elliptical microparticle is situated in a transverse plane to the optical axis. When shifting the ellipsoid from the optical axis, an optical trapping force appears that prevents its displacement, meaning that the particle finds itself in an optical trap on the optical axis.

578-581

Authors: S.V. Karpeev, V.V. Podlipnov, N.A. Ivliev, S.N. Khonina

The possibility of constructing a near-infrared atmospheric optical communication system based on a pair of media converters of a signal format 100Base-TX / 1000BASE-T to format 1000BASE-SX / LX with SFP transceivers DEM-310GT was experimentally investigated. The FAN-OUT TUBING FTB900 SN-Y4 fiber cable connector, coming from the receiving radiation collimator for matching with the DEM-310GT transceiver, was modified. The transmitting radiation collimator is supplemented by a spiral phase plate to form a vortex beam. The influence of atmospheric influence on the data rate is analyzed.

589-595

Authors: Y.V. Vizilter, V.S. Gorbatsevich, A.S. Moiseenko

Facial landmark detection is an important sub-task in solving a number of biometric facial recognition tasks. In face recognition systems, the construction of a biometric template occurs according to a previously aligned (normalized) face image and the normalization stage includes the task of finding facial keypoints. A balance between quality and speed of the facial keypoints detector is important in such a problem. This article proposes a CNN-based one-stage detector of faces and keypoints operating in real time and achieving high quality on a number of well-known test datasets (such as AFLW2000, COFW, Menpo2D). The proposed face and facial landmarks detector is based on the idea of a one-stage SSD object detector, which has established itself as an algorithm that provides high speed and high quality in object detection task. As a basic CNN architecture, we used the ShuffleNet V2 network. An important feature of the proposed algorithm is that the face and facial keypoint detection is done in one CNN forward pass, which can significantly save time at the implementation stage. Also, such multitasking allows one to reduce the percentage of errors in the facial keypoints detection task, which positively affects the final face recognition algorithm quality.

606-617

Authors: I.G. Palchikova, E.S. Smirnov, O.A. Barinova, I.V. Latyshov, V.A. Vasiliev, A.V. Kondakov

We discuss aspects of the use and possibilities provided by three-color colorimeters or digital cameras in problems of detecting small color differences by computer vision methods. The spectral dependence of the total color differences between pairs of visually indiscernible monochromatic stimuli is experimentally revealed. An experimental setup based on the UM-2 monochromator is created for producing a digital atlas of monochromatic stimuli at 1-nm increments. The atlas serves to test the color gamut and color differentiation of cameras. It is experimentally shown that in the visible spectral range a color difference of 3 units is detected by pairs of stimuli that are unevenly distributed across the spectrum and differ in wavelengths from 1 to 6 nm. The capabilities of computer vision are tested on the examples of identifying additional texts during a technical and forensic examination of documents. A new algorithm is developed for finding and quantitatively characterizing color difference of inserts based on a digital image of the inscription. In the algorithm, the objective analysis of the image is divided into a block of color segmentation and that of color tone and color difference assessment. With such an approach, the color segmentation block performs preprocessing functions, making a border map for the classes with different colors for the subsequent calculations. The Otsu method of optimal global threshold transformation is for the first time applied to a problem of image segmentation by color saturation. The trial of the algorithm confirms its efficiency in the solution of expert tasks.

627-635

Authors: A.M. Belov, A.Y. Denisova

Earth remote sensing data fusion is intended to produce images of higher quality than the original ones. However, the fusion impact on further thematic processing remains an open question because fusion methods are mostly used to improve the visual data representation. This article addresses an issue of the effect of fusion with increasing spatial and spectral resolution of data on thematic classification of images using various state-of-the-art classifiers and features extraction methods. In this paper, we use our own algorithm to perform multi-frame image fusion over optical remote sensing images with different spatial and spectral resolutions. For classification, we applied support vector machines and Random Forest algorithms. For features, we used spectral channels, extended attribute profiles and local feature attribute profiles. An experimental study was carried out using model images of four imaging systems. The resulting image had a spatial resolution of 2, 3, 4 and 5 times better than for the original images of each imaging system, respectively. As a result of our studies, it was revealed that for the support vector machines method, fusion was inexpedient since excessive spatial details had a negative effect on the classification. For the Random Forest algorithm, the classification results of a fused image were more accurate than for the original low-resolution images in 90% of cases. For example, for images with the smallest difference in spatial resolution (2 times) from the fusion result, the classification accuracy of the fused image was on average 4% higher. In addition, the results obtained for the Random Forest algorithm with fusion were better than the results for the support vector machines method without fusion. Additionally, it was shown that the classification accuracy of a fused image using the Random Forest method could be increased by an average of 9% due to the use of extended attribute profiles as features. Thus, when using data fusion, it is better to use the Random Forest classifier, whereas using fusion with the support vector machines method is not recommended.

646-652

Authors: A.A. Borodinov

The paper considers a problem of determining the user preferred stops in a public transport recommender system. The effectiveness of using various machine learning methods to solve this problem in a system of personalized recommendations is compared, including a support vector method, a decision tree, a random forest, AdaBoost, a k-nearest neighbors algorithm, and a multi-layer perceptron. The described traditional methods of machine learning are also compared with the method proposed herein and based on an estimate calculation algorithm. The efficiency and the effectiveness of the proposed method are confirmed in the work.

660-664

Authors: W.H. Yang

In order to resist network malicious attacks, this paper briefly introduced the network intrusion detection model and K-means clustering analysis algorithm, improved them, and made a simulation analysis on two clustering analysis algorithms on MATLAB software. The results showed that the improved K-means algorithm could achieve central convergence faster in training, and the mean square deviation of clustering center was smaller than the traditional one in convergence. In the detection of normal and abnormal data, the improved K-means algorithm had higher accuracy and lower false alarm rate and missing report rate. In summary, the improved K-means algorithm can be applied to network intrusion detection.

671-679

Authors: D.A. Shepelev, V.P. Bozhkova, E.I. Ershov, D.P. Nikolaev

This paper considers methods for simulating color underwater images based on real terrestrial images. Underwater image simulation is widely used for developing and testing methods for improving underwater images. A large group of existing methods uses the same deterministic image transformation model ignoring the presence of noise in images. The paper demonstrates that this significantly affects the overall quality of underwater images simulation. It is shown both theoretically and numerically that the accuracy of the signal-to-noise ratio of underwater images simulated using a deterministic transformation decreases with increasing distance to the object. To solve this problem, a new model of image transformation for simulating underwater images based on terrestrial images is proposed, which considers the presence of noise in the image and is compatible with all simulating methods from the group under consideration. The paper presents the results of the simulation based on the existing and proposed models, showing that at long distances, the new results are better consistent with real data.

295-318

Authors: N.L. Kazanskiy, M.A. Butt, S.A. Degtyarev, S.N. Khonina

Optical sensors are widely used in the biomedical, chemical and food industries. They provide high sensitivity to changes in the refractive index of the environment due to a specific distribution of resonances across the field. The sensitivity of the sensor is highly dependent on its material and structure. In this review, we focused on the analysis of silicon waveguides as a promising component for optical sensor miniaturization, and plasmon refractive index sensors without fluorescent labeling. We presented the latest developments of special types of plasmon structures, such as metal-insulator-metal waveguides, and their application in refractive index sensors. We analyzed numerous types of plasmon waveguides, their geometry, materials and manufacturing processes, as well as possible energy losses. A discussion of the spectral characteristics of recently proposed refractive index sensors, with an emphasis on their sensitivity and quality indicators, is an important part of the review.

325-332

Authors: A.S. Abramov, D.A. Evseev, D.I. Sementsov

We investigated optical-range waveguide modes propagating in a semiconductor film sandwiched between two graphene plates. The mode characteristics were shown to depend on the chemical potential of graphene and the film thickness. Based on the numerical analysis, we obtained dispersion relations for the first waveguide modes, frequency dependences of their group and phase velocities, and the distribution of the energy flux density in the structure. We discovered the presence of spectral bands characterized by small phase and negative group velocities of the waveguide modes. The possibility of tuning the waveguide mode by changing the chemical potential of graphene and the thickness of the semiconductor film was established.

343-351

Authors: R.S. Zakirullin

The effect of diffraction on the directional light transmittance of grating optical filters designed for smart windows is studied. The filter has an angular selective light transmission due to two thinfilm gratings formed by "non-transmissive" (absorptive, reflective or scattering) parallel strips on the surfaces of a single or double glazed window. Methods are considered for calculating the optimal slope angle of two filter’s gratings, their relative position on the opposite window surfaces and the widths of the strips of both gratings to minimize light transmission on a specific date and time of day for a given latitude and longitude of the building and azimuth of the window. A method has been developed for calculating the diffraction at the input and output gratings of the filter with regard for a complex movement of the Sun relative to the window and a corresponding change in the incidence angle of the solar beams on the gratings. To evaluate a decrease in the light transmittance of the filter due to diffraction, an additional factor is introduced into the calculation equation. Geometric parameters of the filters and diffraction are calculated. The values of the diffraction factor are 0.9999992 and 0.9999998 for single or double glazed windows, respectively, i.e., the influence of diffraction on the light transmittance of the grating filter can be neglected.

363-374

Authors: Afef Bettaieb, Nabila Filali, Taoufik Filali, Habib Ben Aissia

Image processing is an effective method for characterizing various two-phase gas/liquid flow systems. However, bubbly flows at a high void fraction impose significant challenges such as diverse bubble shapes and sizes, large overlapping bubble clusters occurrence, as well as out-offocus bubbles. This study describes an efficient multi-level image processing algorithm for highly overlapping bubbles recognition. The proposed approach performs mainly in three steps: overlapping bubbles classification, contour segmentation and arcs grouping for bubble reconstruction. In the first step, we classify bubbles in the image into a solitary bubble and overlapping bubbles. The purpose of the second step is overlapping bubbles segmentation. This step is performed in two subsequent steps: at first, we classify bubble clusters into touching and communicating bubbles. Then, the boundaries of communicating bubbles are split into segments based on concave point extraction. The last step in our algorithm addresses segments grouping to merge all contour segments that belong to the same bubble and circle/ellipse fitting to reconstruct the missing part of each bubble. An application of the proposed technique to computer generated and high-speed real air bubble images is used to assess our algorithm. The developed method provides an accurate and computationally effective way for overlapping bubbles segmentation. The accuracy rate of well segmented bubbles we achieved is greater than 90 % in all cases. Moreover, a computation time equal to 12 seconds for a typical image (1 Mpx, 150 overlapping bubbles) is reached.

385-392

Authors: E.A. Shalimova, E.V. Shalnov, A.S. Konushin

Some computer vision tasks become easier with known camera calibration. We propose a method for camera focal length, location and orientation estimation by observing human poses in the scene. Weak requirements to the observed scene make the method applicable to a wide range of scenarios. Our evaluation shows that even being trained only on synthetic dataset, the proposed method outperforms known solution. Our experiments show that using only human poses as the input also allows the proposed method to calibrate dynamic visual sensors.

401-408

Authors: S.V. Sai, A.G. Shoberg

The article proposes a method for controlling the transmission quality of fine image details in the JPEG2000 standard based on an automatic adjustment of the quantization parameters of discrete wavelet transform (DWT) coefficients. An algorithm for setting up quantization scale parameters for different transformation sub-ranges of the DWT coefficients depending on given (permissible) distortions is described. For an objective assessment of image quality, numerical measures of fine detail distortion in a normalized N-CIELAB colorimetric system are used, according to which an analysis of their structural features is performed. Results of the experimental studies of the analysis of image quality and compression efficiency depending on the quantization parameters utilized in the developed adaptive compression algorithm are presented. Results of evaluating the algorithm performance that can be used for practical applications in multimedia applications are also presented.

420-426

Authors: P.V. Gulyaev

The article discusses techniques for the precise positioning of a probe tip on the surface under study in scanning probe microscopy. The precise tip positioning is of high importance when studying the same sample with various microscopes, probes, in various environments and conditions. It is offered that the precise positioning should be done by applying on the surface equidistant reference marks with a nanoindentor along a straight line. The article describes a procedure for shifting the microscope’s field of view for detecting the reference marks and subsequently moving the probe along the marks line. A reference marks recognition algorithm is presented. The algorithm is based on the image correlation analysis using a master mark image and the subsequent extraction of special points (at local maxima of the correlation coefficient). A descriptor of mutual distances between the special points and the Pearson coefficient (аs a quantitative criterion of recognition) are used for recognition. Three types of reference marks are considered: a circle, a circle with gradient filling, and a circle with continuous filling. It is shown that the number of the allocated special points can be significantly larger than that of the reference marks. A method for filtering special points is offered. The method is based on the threshold filtering of special points using values of the correlation coefficient. The filtering consists in excluding those special points in whose neighborhood the difference of the maximum and minimum of the correlation coefficient does not exceed a preset threshold. Results confirming the effectiveness of the offered filtering method are presented. Recommendations for choosing the parameters of the recognition and filtering algorithms are given.

436-440

Authors: M.V. Gashnikov

Parameterized interpolation algorithms are adapted to fusion of multidimensional signals of various resolutions. Interpolating functions, switching rules for them and local features are specified, based on which the interpolating function is selected at each point of the signal. Parameterized interpolation algorithms are optimized based on minimizing the interpolation error. The recurrent interpolator optimization scheme is considered for the situation of inaccessibility of interpolated samples at the stage of setting up the interpolation procedure. Computational experiments are carried out to study the proposed interpolators for fusion of real multidimensional signals of various types. It is experimentally confirmed that the use of parameterized interpolators allows one to increase the accuracy of signal fusion.

446-453

Authors: Ye.V. Goshin, A.P. Kotov

In this paper, we study methods for determining parameters of camera movement from a set of corresponding points. Unlike the traditional approach, the corresponding points in this paper are not used to determine the fundamental matrix, but directly to determine motion parameters. In addition, in this work, we use a multi-angle image formation model based on the representation of three-dimensional images and motion parameters in the form of quaternions. We propose method for determining motion parameters, including the selection of the most noise-free matches using the RANSAC method. The study presents results of an experiment on the “Middlebury” and “ETH3D” test kits, which contains a set of images with known values of the motion parameters. Using a program written in Python, a comparative experiment was conducted to evaluate the accuracy and reliability of the estimates obtained using the proposed method under conditions of a small number of corresponding points and a shallow depth of the scene. In the course of experimental studies, it was shown that under the above-described conditions, the reliability of parameter determination using the proposed method significantly exceeds the reliability of traditional methods for estimating motion parameters based on the calculation of the fundamental matrix.

466-475

Authors: Y.A. Mostovoy , V.A. Berdnikov

When solving a given common task, a swarm of moving objects coordinates the state of its individual members. When planning swarm operations, there is a need to take into account the possibility of its operational regrouping, since at the time of planning the exact purpose of the swarm operation may be either not yet defined, or secret, or determined by a number of random circumstances. At the same time, the swarm resources are insufficient for one-time coverage of all possible targets in a given service area. Therefore, the execution of the operation of the swarm begins to resolve the aforementioned uncertainties. In this case, the operation time can be significantly reduced. This problem is solved by methods of the percolation theory. A concept of programmable percolation of the service area, which is implemented in two phases, is introduced. The value of the concentration of objects in a basic stochastic swarm is obtained numerically -- using the results of statistical modeling of two-phase operations -- and analytically, providing a minimum total cost of the two-phase operation. The synergy of information interaction between the swarm objects when implementing a programmable percolation path is analyzed.

482-487

Authors: A.D. Bragin, V.G. Spitsyn

Electroencephalography is a widespread method to record brain signals with the use of electrodes located on the surface of the head. This method of recording the brain activity has become popular because it is relatively cheap, compact, and does not require implanting the electrodes directly into the brain. The article is devoted to a problem of recognition of motor imagery by electroencephalogram signals. The nature of such signals is complex. Characteristics of electroencephalograms are individual for every person, also depending on their age and mental state, as well as the presence of noise and interference. The multitude of these parameters should be taken into account when analyzing encephalograms. Artificial neural networks are a good tool for solving this class of problems. Their application allows combining the tasks of extracting, selecting and classifying features in one signal processing unit. Electroencephalograms are time signals and we note that Gramian Angular Fields and Markov Transition Field transforms are used to represent time series in the form of images. The article shows the possibility of using the Gramian Angular Fields and Markov Transition Field transformations of the electroencephalogram (EEG) signal for motor imagery recognition using examples of imaginary movements with the right and left hand, also studying the effect of the resolution of Gramian Angular Fields and Markov Transition Field images on the classification accuracy. The best classification accuracy of the EEG signal into the motion and state-of-rest classes is about 99%. In future, the research results can be applied in constructing the brain-computer interface.

145-154

Authors: V.V. Kotlyar, A.A. Kovalev, A.V. Volyar

An optical vortex passed through an arbitrary aperture (with the vortex center found within the aperture) or shifted from the optical axis of an arbitrary axisymmetric carrier beam is shown to conserve the integer topological charge (TC). If the beam contains a finite number of off-axis optical vortices with different TCs of the same sign, the resulting TC of the beam is shown to be equal to the sum of all constituent TCs. For a coaxial superposition of a finite number of the Laguerre-Gaussian modes (n, 0), the resulting TC equals that of the mode with the highest TC (including sign). If the highest positive and negative TCs of the constituent modes are equal in magnitude, then TC of the superposition is equal to that of the mode with the larger (in absolute value) weight coefficient. If both weight coefficients are the same, the resulting TC equals zero. For a coaxial superposition of two different-amplitude Gaussian vortices, the resulting TC equals that of the constituent vortex with the larger absolute value of the weight coefficient amplitude, irrespective of the relation between the individual TCs.

167-176

Authors: E.K. Bashkirov

In this paper, we investigate the entanglement between two two-level atoms non-resonantly in-teracting with a thermal field of a lossless one-mode resonator via degenerate two-photon transi-tions. On the basis of the exact solution of the time-dependent density matrix we calculate the negativity as a measure of atomic entanglement. We show that for separable initial atomic states a slight atom-field detuning may generate the high amount of atom-atom entanglement. The re-sults also show that for non-resonant atom-field interaction the entanglement induced by nonlin-ear two-photon interaction is smaller than that induced by one-photon interaction in contrast to the resonant interaction situation. For a Bell-type entangled initial atomic state we obtain that if the detuning increases, there is an appreciable decrease in the amplitudes of the negativity oscilla-tions. The results also show that elimination of the sudden death of entanglement for non-resonant two-photon atom-field interaction may take place.

183-188

Authors: Yu.S. Dadoenkova, I.O. Zolotovsky, I.S. Panyaev, D.G. Sannikov

The amplification and generation of optical TE waves arising on a lattice formed by a space charge wave in a planar waveguide based on a donor-doped semiconductor (gallium arsenide) n-GaAs are considered. The region of interaction is limited by contacts with a constant electric field applied between them, which, while operating in the Gunn oscillations suppression mode, gener-ates a small-signal periodic inhomogeneity. Reflection and transmission regimes for same-index TE modes propagating in the waveguide structure are investigated depending on the phase mis-match and the pump level. It is shown that even with a relatively small modulation depth of the di-electric constant (about 10–5), under conditions of high optical pumping (with a gain of about 150 cm–1) and a corresponding detuning from phase matching there is the possibility of not only ampli-fying the direct and backward (reflected) optical modes, but also their generation. The advantage of the considered effects is that they enable flexible control of parameters of the dynamic lattice. The results obtained can be used to create semiconductor laser generators based on the interaction of the optical modes and a space charge wave.

195-202

Authors: V.M. Vladimirov, V.A. Yukseev, E.G. Lapukhin

In this work, we propose a high-resolution optical system for Earth remote sensing, operating at 200-3300-nm wavelengths and providing a 4º field of view. Parameters of the system's structural elements are calculated and presented. Dot charts of the spots of confusion for the center, intermediate zones, and the edge of the field are considered. Over most of the operating wavelengths, the optical system is shown to be diffraction-limited, which provides attaining the highest possible spatial resolution. This system is considered as a tool for monitoring of the Earth's surface and collection of information in the ultra-violet, visible and near infrared spectral ranges (200-3300 nanometers).

209-213

Authors: Hanh Hong Mai

This study presents a new construction design of a distributed Bragg reflector (DBR) filter and a Fabry–Pérot (FP) filter by using needle technique as a synthesis method. The optimized DBR and FP filters having a proper number of layers with controlling thickness TiO2/SiO2 are utilized to transmit only a certain narrow band of wavelengths while blocking the others. As a proof of concept, the filters are designed to selectively transmit only a very narrow band of wavelength at 780 nm which is the near infrared (NIR) fluorescent emission from Alexa Fluor 750 dye. The obtained results show that the optimized filters represent advanced spectral performance which can be used to improve the sensitivity and the imaging contrast in fluorescence microscopy.

219-228

Authors: D.V. Nesterenko

The resonance characteristics of the Fabry-Pérot resonator modes supported by metal/dielectric/metal planar structures are studied in the case of absorbing media for near-to-normal light incidence. Approximations based on rigorous solution and field-transfer model for the field and resonance line shapes in spectra are attributed to the class of Fano and Lorentz resonances. The analytical expressions are obtained for the propagation constant and field enhancement of the mode, width, height and slope of resonance line shapes in spectra as functions of structural parameters. With estimation of field characteristics of the fabricated loss structures based on aluminum and quartz, the peaks in the transmission spectra can be attributed to the excitation of Fabry-Pérot modes. Fundamental characterization of Fabry-Pérot resonances may find applications in optical processing and sensing.

236-243

Authors: B.V. Faizov, V.I. Shakhuro, V.V. Sanzharov, A.S. Konushin

The paper studies the possibility of using neural networks for the classification of objects that are few or absent at all in the training set. The task is illustrated by the example of classification of rare traffic signs. We consider neural networks trained using a contrastive loss function and its modifications, also we use different methods for generating synthetic samples for classification problems. As a basic method, the indexing of classes using neural network features is used. A comparison is made of classifiers trained with three different types of synthetic samples and their mixtures with real data. We propose a method of classification of rare traffic signs using a neural network discriminator of rare and frequent signs. The experimental evaluation shows that the proposed method allows rare traffic signs to be classified without significant loss of frequent sign classification quality.

250-258

Authors: N.Yu. Ilyasova, N.S. Demin, A.S. Shirokanev, A.V. Kupriyanov, E.A. Zamytskiy

The paper proposes a method for selection the region of diabetic macular edema in fundus images using OCT data. The relevance of the work is due to the need to create support systems for laser coagulation to increase its effectiveness. The proposed algorithm is based on a set of image segmentation methods, as well as searching for specific points and compiling their descriptors. The Canny method is used to find the boundary between the vitreous body and the retina in OCT images. The segmentation method, based on the Kruskal algorithm for constructing the minimum spanning tree of a weighted connected undirected graph, is used to select the retina to the pigment layer in the image. Using the results of segmentation, a map of the thickness of the retina of the eye and its deviation from the norm were constructed. In the course of the research, the optimal parameter values were selected in the Canny and graph segmentation algorithms, which allow to achieve a segmentation error of 5 %. SIFT, SURF, and AKAZE methods were considered for super-imposing calculated maps of the retina thickness and its deviation from the norm on the fundus image. In cases where a picture from the fundus camera of the OCT apparatus is provided along with OCT data, using the SURF method, it is possible to accurately combine with the fundus image.

266-273

Authors: Yu.D. Agafonova, A.V. Gaidel, P.M. Zelter, A.V. Kapishnikov

We compare approaches for the automatic detection of pathological changes in brain MRI images that are visible to the naked eye. We analyse multi-stage approaches based on deep learning and threshold processing. A convolutional neural network was formed, a classifier was built based on the use of an ensemble of decision trees, and an algorithm was created for multi-stage image processing. Because of experimental studies, it was found that the most effective method for recognizing images of magnetic resonance imaging is an approach based on an ensemble of decision trees. With its help, 95 % of the images from the test sample were classified correctly. At the same time, using the convolutional neural network, it was possible to classify correctly all images containing the area of pathological changes. The data obtained can be used in practice for the diagnosis of brain diseases, for automating the processing of a large number of studies of magnetic resonance imaging.

282-289

Authors: I.M. Kulikovsvkikh

Previous research in deep learning indicates that iterations of the gradient descent, over separable data converge toward the L2 maximum margin solution. Even in the absence of explicit regularization, the decision boundary still changes even if the classification error on training is equal to zero. This feature of the so-called “implicit regularization” allows gradient methods to use more aggressive learning rates that result in substantial computational savings. However, even if the gradient descent method generalizes well, going toward the optimal solution, the rate of convergence to this solution is much slower than the rate of convergence of a loss function itself with a fixed step size. The present study puts forward the generalized logistic loss function that involves the optimization of hyperparameters, which results in a faster convergence rate while keeping the same regret bound as the gradient descent method. The results of computational experiments on MNIST and Fashion MNIST benchmark datasets for image classification proved the viability of the proposed approach to reducing computational costs and outlined directions for future research.