Home – Home – Jornals – Avtometriya 2021 number 2

An approach to the mathematical description of signals and jamming as multidimensional spatiotemporal signals is considered. The algorithm for signal processing in the presence of internal noise and active noise jamming is proposed and described. The algorithm is applicable for the cases of receiving a signal with completely known parameters and a signal with a random initial phase. The variant of a linear digital antenna array is investigated.

For the classification of fragments of a hyperspectral image, a preliminary transformation of its spectral features to the principal components and subsequent recognition using a convolutional neural network trained on a sample composed of fragments of this image are demonstrated to be very effective. High percentages of the correct classification are obtained with a large-format hyperspectral image, despite the fact that some of the classes into which the hyperspectral image is divided are very close to each other and, accordingly, are difficult to distinguish by hyperspectra. The dependences of the correct classification on the size of the fragments, from which the training and validation samples are composed, and on the parameters of the convolutional neural network are investigated.

A new approach to calculating the characteristics of the gradient of a grayscale image as an array of features of the object of interest is considered. It is proposed to design a model of a reconfigurable computing environment that can simultaneously process each pixel of the source image in a parallel mode and form an array with gradient characteristics. Owing to the architectural principles of model construction, the gradient calculation is performed in one clock cycle of the elementary calculator of the reconfigurable computing environment.

A method for visualizing multi-volume data and functionally defined surfaces using graphics processing units is proposed. The method provides visualization of a large number of volumes, a set of translucent and functionally defined objects, and complex translucent volumes, including volume intersections in constructive solid modeling. Rendering different volumes at the same time is a more complex problem than rendering a single volume because intersecting and mixing operations are required. Functionally defined surfaces are well suited for embedding external objects into volumes.

A new method is proposed for testing a hypothesis on independence of two-dimensional random variables. The method under consideration is based on the use of a nonparametric pattern recognition algorithm that meets the maximum likelihood criterion. In contrast to the traditional problem statement, there is no training sample a priori. The initial information is represented by statistical data, which are the values of two-dimensional random variables. The distribution laws of random variables in classes are estimated according to the initial statistical data for the conditions of their dependence and independence. In choosing the bandwidths for nonparametric probability density estimation, the maximum of the likelihood functions is used as a criterion. Under these conditions, the estimates of the probability of an error in pattern recognition in classes are calculated. Based on the minimum value of the estimate of the pattern recognition error probability, a decision is made whether the random variables are dependent or independent. The effectiveness of the developed technique is confirmed by the results of computational experiments with testing the hypothesis on independence or linear dependence of two-dimensional random variables.

Statistical characteristics of a narrow-band signal in the presence of modulating noise are considered, including the case of determination of the univariate density of the signal distribution. The relationship between the characteristic function of the instantaneous signal values and the probability distribution density (PDD) envelope with a uniform phase distribution law is determined. Expressions describing the PDD of the instantaneous values of the signal and additive noise under the action of modulating noise are obtained. Cases without and with a functional relationship between the amplitude and phase fluctuations are analyzed. With the phase distribution law other than uniform on the interval [0, 2π], the signal is demonstrated to be a non-stationary random function tending to stationary as the depth of phase distortions increases. With a uniform phase distribution, the signal is a stationary random function. In the case of functionally connected amplitude-phase distortions, the PDD signal is described through the statistical characteristics of its envelope.

The redundancy of universal encoding of Markov sources defined by transition probability matrices with a fixed number of different strings by non-uniform symbols is found. As a corollary, redundancy estimates are obtained for Markov sources with memory s and Markov sources of Miley defined by a graph. The rate of redundancy decreasing is determined depending on the graph characteristics, the encoded block length, and the channel bandwidth.

For analytical work with the electronic structure of atomic systems and for the purposes of training specialists in the field of spectroscopy, scientific visualization of levels and transitions in an atomic system is important, which usually implemented in the form of Grotrian diagrams. Currently, such diagrams can be generated programmatically from spectroscopic databases. Unlike printed graphic images, the graphics processed and presented in a machine way may have interactivity and dynamics properties, which significantly affects the efficiency of specialist's work with it. It may be assumed that the possibilities of machine visualization can give rise to new types of graphical representations of the electronic structure, which were not used previously because of their laboriousness or other reasons. In this paper, a new type of diagrams of the electronic structure of atomic systems, called quantograms by the authors, is proposed. It is focused on computer generation combined with an interactive dynamic presentation of the results. The idea of the new diagrammatic technique, its main features, the experience of software implementation, and working with quantograms automatically generated from a spectral database are described.

The challenges of the x-ray inspection of steel objects with a thickness of 100 mm and more are discussed. In the classical readout method, semiconductor photosensors are located directly on scintillators. As a result, the background interaction of gamma radiation with the photosensor material is responsible for image noise and causes negative effects in the readout electronics. The method proposed to design registration system with using the wavelength shifters is free from disadvantages of the classical method. This method allows one to achieve smaller noise of the image as compared to the classical method, avoid the gamma radiation impact on electronics, and enhance the inspection efficiency. Various ways of implementation of the proposed method are experimentally studied, and optimal methods are chosen. Based on the investigations performed, a design of the registration element is proposed, and results of measuring its parameters are reported.

Partial discharges (PDs) appear in defects in the basic insulation of high-voltage electrical equipment (HVE) elements, which are inaccessible for direct measurements, but are the main diagnostic sign. The aim of this study is to build a correct diagnostic model of a defect in HVE insulation as a source of PD series and to study the problem of differential diagnosis of individual defects based on the interpretation of the integral picture of PD as a set of time instants of individual elementary PDs available for measurement. The idea of a mathematical problem of covering a set is used for diagnostics. A model example is considered.

The vibroactivity of micro-cryogenic machines is an important factor determining the resource characteristics of photodetector arrays (PDAs) operating in the IR range. The article describes the amplitude-frequency characteristics of the vibration activity of PDAs with different service lives produced by different manufacturers. The vibroactivity is measured by a strain sensor rigidly connected with the PDA base. The influence of centrifugal and compression forces on the PDA vibroactivity and the possibility of their mutual compensation are analyzed.

In the present paper, we propose a statistical algorithm of radio pulse energy estimation in measuring the superconducting qubit-resonator system. The algorithm uses complete sufficient statistics, and it does not require a priori information about the level of additive white Gaussian noise and the radio pulse amplitude, initial phase, and time position. The obtained estimates are efficient and consistent. The algorithm performance is evaluated by computer simulation methods.

A quasi-distributed resistive sensor with a tree structure is proposed. The sensor makes it possible to measure the distribution of physical quantities when switching measuring circuits only to external terminals. It is shown that the use of a tree structure provides high accuracy with low computational costs.

A new version of the reflective interferometer for S-polarized light consists of a thin metal film placed in front of a multilayer dielectric interferometer. The appearance of narrow extrema (maxima or minima) in the spectral or angular dependence of the reflection coefficient depends on the location of the metal film in the node or the antinode of the standing wave reflected from the interferometer.