=Paper=
{{Paper
|id=Vol-3171/paper84
|storemode=property
|title=Information-Extreme Machine Learning of an On-board Ground Object Recognition System with a Choice of a Base Recognition Class
|pdfUrl=https://ceur-ws.org/Vol-3171/paper84.pdf
|volume=Vol-3171
|authors=Igor Naumenko,Vladyslav Piatachenko,Mykyta Myronenko,Taras Savchenko
|dblpUrl=https://dblp.org/rec/conf/colins/NaumenkoPMS22
}}
==Information-Extreme Machine Learning of an On-board Ground Object Recognition System with a Choice of a Base Recognition Class==
https://ceur-ws.org/Vol-3171/paper84.pdf
Information-Extreme Machine Learning of an On-board Ground
Object Recognition System with a Choice of a Base Recognition
Class
Igor Naumenko1, Vladyslav Piatachenko2, Mykyta Myronenko2 and Taras Savchenko2
1
Scientific-research center of missile troops and artillery, Gerasim Kondratyev st, 165, Sumy, 400021, Ukraine
2
Sumy State University, Rymskogo-Korsakova st. 2, Sumy, 40007, Ukraine
Abstract
The aim of the work is to increase the functional efficiency of machine learning of the on-
board system for recognizing land-based natural and infrastructural objects by optimizing the
incoming mathematical description. Within the framework of information-extreme intelligent
data analysis technology, which is based on maximizing the amount of information in the
machine learning process, a method of information-extreme synthesis of onboard recognition
system has been developed. Within the framework of a functional approach to modeling
cognitive processes of natural intelligence, a categorical functional model of information-
extreme machine learning of an on-board recognition system with an automatic selection of
the base class of recognition is proposed. Based on the proposed category model, an algorithm
of information-extreme machine learning has been developed and programmatically
implemented. The dependence of the machine learning functional efficiency on the choice of
the base recognition class is experimentally investigated, in relation to which a system of
control tolerances for recognition features is determined in the machine learning process. As a
criterion for optimizing the parameters of machine learning, a modified information measure
of Kullback is used, which is considered as a functional from the exact characteristics of
classification solutions. Based on the machine learning results, the decision rules ensured high
accuracy of digital image segmentation in the region.
Keywords 1
Information-extreme machine learning, onboard recognition system, digital image of the
region, information criterion, optimization
1. Introduction
Use The task of recognizing ground objects by the on-board system of an unmanned aviation
complex divided into two stages: searching on an electronic map of the region of interest areas in which
there is the greatest probability of finding the object being sought, and direct recognition of the object
in the area of interest. Various natural areas of the region and infrastructural constructions, which also
include highways and railways, bridges, airports, etc., can be areas of interest when recognizing small-
scale terrestrial objects. Modern experience in the use of unmanned aerial vehicles shows that the search
and recognition of ground objects is carried out mainly in an interactive mode, in which the onboard
recognition system (ORS) performs the functions of the region a digital image translator to the ground
control point. At the same time, there is a trend towards the development of autonomous ORS, which
allows to expand their functionality to solve a wide range of problems and increases the cyber security
of unmanned aerial vehicles. As a promising way of information synthesis of autonomous onboard
systems is the application of machine learning ideas, methods and pattern recognition. The reliability
COLINS-2022: 6th International Conference on Computational Linguistics and Intelligent Systems, May 12-13, 2022, Gliwice, Poland
EMAIL: 790905@ukr.net (I. Naumenko); vl.piatachenko@cs.sumdu.edu.ua (V. Piatachenko); nikitam1996@ukr.net (M. Myronenko);
taras.savchenko01@gmail.com (T. Savchenko)
ORCID: 0000-0003-2845-9246 (I. Naumenko); 0000-0002-7464-3119 (V.Piatachenko); 0000-0001-5005-1672 (M. Myronenko); 0000-
0002-9557-073X (T. Savchenko)
Β©οΈ 2022 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
of the search for areas of interest in the electronic image of the region depends mainly on two main
reasons:
β’ the adequacy of the input mathematical description of the on-board system for identifying
digital image frames of the region to real conditions;
β’ functional efficiency of ORS machine learning.
Since the onboard system of a modern unmanned aerial vehicle is characterized by high computing
power, and available on-board video cameras, thermal imagers and other surveillance devices have a
high resolution, now there are all the technical conditions for processing and operational analysis of
digital images. But the main deterrent to the introduction of autonomous onboard ORS are scientific
and methodological complications associated primarily with arbitrary initial conditions for the
formation of the ground object image, the intersection in the space of recognition classes features and
a large amount of data.
The paper considers the problem of information-extreme synthesis of the learnable on-board
recognition system, which identifies frames of the digital image of the region in order to determine the
areas of interest in which the vehicle may be located.
2. Related Works
The most common for the formation of the input mathematical description of the BSR are descriptor
methods that can distinguish the contours of ground objects [1 - 3]. The main disadvantages of this
approach are the lack of information about the recognition features, as they do not take into account the
local design features and external characteristics of the vehicle, and low efficiency of recognition.
Higher efficiency of object recognition by contour is achieved by applying the method of proportional
coefficients [4]. The most promising is detection based on scanning the entire image of a small ground
object [5 - 7]. In [8, 9] the possibility of using neuro-like structures to solve the problems of the ORS
autonomous functioning is considered. The main disadvantage of this approach is the sensitivity of
artificial neural networks to the multidimensionality of the recognition features space and the
recognition classes alphabet. n addition, since the recognition classes in practice intersect in the space
of features, the works [10] are known, in which fuzzy neural networks are solved using the apparatus
of fuzzy logic. The article [11] proposes the method of solving the problem of feature space
multidimensionality via dividing the input data on blocks and submit them to multilayer extractors of
neural networks. Encoded data from extractor output is submitted to megaclassifier. Such approach
leads to data loss and affects accuracy of the classification decisions. It is known that the greatest
efficiency of classification decisions is characterized by methods of pattern recognition, which are able
to build decision rules within the geometric approach [12]. Within this approach as a promising
direction is the use of ideas and methods of so-called information-extreme intellectual technology (IEI-
technology), which is based on the implementation of the maximizing information principle in the
process of machine learning system [13 - 15]. In [16], the segmentation of the region digital image was
considered within the framework of IEI-technology, but the problem of choosing the basic recognition
class and its influence on the functional efficiency of ORS machine learning was not investigated.
The purpose of the work is to increase the functional efficiency of the ORS of natural and
infrastructural ground objects by choosing a base recognition class, in relation to which a system of
control tolerances for recognition signs.
3. Methods and Materials
Consider the formalized formulation of the ORS information-extreme synthesis problem of
terrestrial objects capable of learning with the automatic selection of the basic recognition class. Let the
alphabet of {ππ 0
|π = Μ
Μ
Μ
Μ
Μ
Μ
1, π} recognition classes be formed, which characterize the frames of the region
digital image obtained by aerial photography. For each recognition class, a three-dimensional learning
(π) (π)
matrix βπ¦π,π β of brightness is formed, in which row {π¦π,π |π = Μ
Μ
Μ
Μ
Μ
1, π}, where π is the number of
recognition features, is a structured vector of features of the corresponding recogniton class, and the
(π)
matrix column is a random training sample {π¦π,π |π = Μ
Μ
Μ
Μ
Μ
1, π} of the π-th feature with volume n.
It is known that one of the characteristic features of the methods of IEI-technology is the
transformation of the input training matrix π into a working binary matrix π, which adapts in the process
of machine learning to the maximum full probability of correct classification decisions. Therefore, the
Hamming binary space is given a vector of functioning parameters that affect the functional efficiency
0
of ORS machine learning to recognize structured feature vectors, for example, recognition class ππ :
ππ =< π₯π , ππ , πΏ > , (1)
where π₯π is the vector of features averaged over the training matrix, the vertex of which determines
π
the center of the hyperspherical container of recognition class ππ in the Hemming bemming space; ππ
π
radius of the hyperspherical container of recognition class ππ , which is restored in the radial basis of
the recognition features space; πΏ β parameter, the value of which is equal to half the symmetric field of
control tolerances for recognition features.
The parameters of the system, which will be called the parameters of machine learning, are subject
to appropriate restrictions:
β’ the recognition container radius values range ππ π
is given by inequality ππ < π(π₯π β π₯π ),
where π(π₯π β π₯π ) β is the center-to-center distance between the averaged feature vector π₯π and the
nearest corresponding feature vector π₯π of the neighboring class πππ ;
β’ the parameter values range ο€ is given by inequality πΏ < πΏπ» /2, where πΏπ» is the normalized
field of tolerances for recognition features;
Figure 1 shows a bilateral symmetrical field of control tolerances.
Figure 1: Recognition tolerance fields
In Figure 1 the following designations are accepted: π΄0,π β nominal (average) value of sign π¦π ;
π΄π»,π , π΄π΅,π β lower and upper normalized tolerances for feature π¦π ; π΄π»πΎ,π , π΄π΅πΎ,π β lower and upper control
tolerances for sign π¦π ; πΏ β parameter equal to half of the symmetric field of control tolerances.
It is necessary to find the basic recognition class for a given alphabet, in relation to which the system
of control tolerances for recognition features is determined, and in the process of ORS machine learning
to optimize machine learning parameters (1), which provide the maximum value of information
criterion in the working:
1 (π)
πΈΜ
β = π βππ=1 πππ₯ πΈπ , (2)
πΊπΈ β©{π}
(π)
where πΈπ is the value of the information criterion calculated at the π-th step of machine learning;
πΊπΈ β working area of the information criterion calculation; {π} βmany machine learning steps.
Thus, the task of information synthesis of the learnable recognition system is to optimize the
parameters of its machine learning by approaching the global maximum of the information criterion (2)
to its maximum limit value.
The development of information-extreme machine learning methods is carried out within the
framework of a functional approach to modeling cognitive processes by natural intelligence. Within the
framework of this approach, we represent the categorical functional model of information-extreme
machine learning BSR in the form of a directed graph of mapping sets to each other with the help of
machine learning operators involved in the process. The input mathematical description of the category
model is given in the form of a structure
πΌπ΅ =< πΊ, π, πΊ, πΎ, π, π; π1 , π2 >, (3)
where πΊ is the set of factors that affect the BSR; π β a set of points in time for reading information;
πΊ β space of recognition signs; πΎβ the region digital image frames set; π β input training matrix of
image frame pixels brightness; πβ working binary training matrix; π1 β operator formation of the input
training matrix π; π2β operator for converting the matrix π into a working binary matrixπ.
The categorical functional model of the SBR information-extreme machine learning with automatic
determination of the basic recognition class is shown in Figure 5.
Figure 2: Categorical functional model of ORS machine learning
In Figure 2 Cartesian product πΊ Γ π Γ πΊ Γ πΎ specifies the source of information. The term set Π of
the information criterion values is common to all circuits for optimizing the parameters of machine
learning. The operator π: πΈ β β Μ |π| builds at each step of the machine learning partition β Μ |π| , which is
displayed by the operator π on the distribution of binary feature vectors. Next, the operator π: π β πΌ |π| ,
(π)
where πΌ |π| is the set π of statistical hypotheses, tests the basic statistical hypothesis πΎ1 : π₯π,π β ππ π
. The
2
operator πΎ determines the set of the classification solution exact characteristics, where π = π , and the
operator π calculates the set of values πΈ of the information optimization criterion, which is a functional
of the exact characteristics. The contour of control tolerances optimization for recognition features is
closed through the term set π· the elements of which are the values of the system of control tolerances
for recognition features. The contour, which includes the set π΄ β the ordered alphabet of recognition
classes, automatically determines the basic recognition class by searching, which provides the
maximum value of the information criterion for optimizing the parameters of machine learning (2).
Operator π’ regulates the process of machine learning.
Within the framework of the functional approach to modeling of cognitive processes of classification
decisions making the categorical functional model (Figure 2) is considered as the generalized structural
scheme of information-extreme machine learning algorithm.
4. Experiment
According to the categorical model (Figure 2), the machine learning algorithm with optimization of
the control tolerances system for recognition features was implemented in the form of a two-cycle
procedure for finding the global maximum of information criterion (2) in the working area of its
function.
1 (π)
πΏ β = πππ πππ₯ [π βπ
π=1 πππ₯ πΈπ ], (4)
πΊπΏ πΊπΈπ β©{π}
where πΊπΏ is the allowable range of the control tolerances field parameter πΏ values for recognition
features.
Optimization of control tolerances for recognition features was carried out according to a parallel-
sequential scheme. The extreme values of machine learning parameters obtained in the process of
parallel optimization are quasi-optimal, because they changed at each step of learning by the same
amount for all features simultaneously. To increase the functional efficiency of ORS, it is advisable to
implement a machine learning algorithm with consistent optimization of control tolerances. Thus the
control tolerances received at a stage of parallel optimization were accepted as starting at consecutive
optimization which can be carried out, for example, by iterative procedure of search of a global
maximum of an information criterion in the form [10]
π=1
πΏ
β 1 (π)
πΏπΎ,π = πππ β {πππ₯ [π βπ
π=1 πππ₯ πΈπ (ππ )]},π = 1, π, (5)
πΊπΏπ πΊπΈπ β©{π}
where πΏ is the number of the optimization sequential procedure runs of control tolerances due to
suboptimal starting values of control tolerances for all features; β β symbol of the repeat operation.
Machine learning ORS with parallel-sequential optimization of control tolerances allows to increase
the reliability of classification solutions and at the same time significantly increases the efficiency of
machine learning, because the search for a global maximum criterion is carried out only in the work
area to determine its function.
As an information criterion for optimizing the parameters of machine learning, a modified Kullback
measure was considered, which for two-alternative solutions with a priori equally probable hypotheses
has the form
(π) (π)
(π) (π) (π) 1+[π·1,π (π)βπ½π (π)]+10βπ
πΈπ = [π·1,π (π) β π½π (π))] Γ πππ2 [ (π) (π) ], (6)
1β[π·1,π (π)βπ½π (π)]+10βπ
(π)
where π·1,π (π) is the first reliability, which characterizes the probability of the recognition features
π (π)
vector correct classification ππ ; π½π (π) is an error of the second kind, which characterizes the
π
erroneous assignment to class ππ of the vector of features of the nearest neighboring class; π β remote
measure, which determines the radii of hyperspherical containers of recognition classes, built in the
radial basis of Hamming binary space; 10βπ β a sufficiently small number, which is entered to avoid
division by zero (the value of r in practice is selected from the interval 1 < π β€ 3).
Since the information criterion is a functional of the exact characteristics, it is necessary to use their
estimates for a representative volume of the training sample:
(π) (π)
(π) πΎ1,π (π) (π) πΎ4,π (π)
π·1,π (π) =
ππππ
; π·2,π (π) =
ππππ
(7)
(π)
where πΎ1,π (π) is the number of events that indicate the belonging of βtheirβ implementations of
(π)
π
recognition class ππ ; πΎ4,π (π) β the number of events that mean the non-belonging of βforeignβ
π
implementations of recognition class ππ ; ππππ β the minimum size of a representative training sample,
which is determined by the method proposed in the work [10].
After substituting the corresponding notation (6) in expression (5) we obtain a working formula for
calculating within the IEI-technology criterion for optimizing for optimizing the parameters of machine
π
learning to recognize the vectors of the signs of classππ :
(π) (π)
(π) 1 (π) (π) 1+[πΎ1,π βπΎ2,π ]+10βπ
πΈπ = π [πΎ1,π β πΎ2,π ] Γ πππ2 [ (π) (π) ] (8)
πππ 1β[πΎ1,π βπΎ2,π ]+10βπ
The normalized form of criterion (7) has the form
(π)
(π) πΈπΎπ
πΈπΎ,π = (π) (9)
πΈπΎπππ₯
(π) (π) (π)
where πΈπΎπππ₯ is the value of criterion (7) for πΎ1,π (π) = πΎ2,π (π) = ππππ .
(π) (π)
The calculation of coefficients πΎ1,π and πΎ2,π was carried out according to the procedures
(π) π (π)
πππ₯π β ππ π‘βππ πΎ1 (π): = πΎ1 (π β 1) + 1; πππ₯π β πΡπ π‘βππ πΎ2 (π): = πΎ2 (π β 1) + 1.
In this case, the assignment, for example, implementation π₯ (π) to the recognition class ππ
π
is carried
out according to the rule:
1. the code distance π[π₯π β π₯ (π) ] is calculated;
2. comparison: if π[π₯π β π₯ (π) ] β€ ππ , then π₯ (π) β ππ π
, else β π₯ (π) β ππ
π
;
According to the optimal geometric parameters of the recognition classes containers obtained in the
machine learning process, decision rules are built for the identification of frames of the digital image
of the region during the operation of the ORS in the examination mode.
For hyperspherical containers of recognition classes, the decision rules have the form [11]
π
(βππ β β|π| )(βπ₯ (π) β β|π| )[ππ(ππ > 0)& (ππ = πππ₯ {ππ })
{π}
π‘βππ π₯ (π) β ππ
π
πππ π π₯ (π) β ππ
π
], (10)
(π) (π)
where π₯ is a recognizable vector; ππ β membership function of vector π₯ of the container of the
π
recognition class ππ .
π
In expression (10), the membership function for a hyperspherical container of recognition class ππ
is determined by the formula
β βπ₯ (π) )
π(π₯π
ππ = 1 β β
ππ
, (11)
β β
where π₯π is the optimal averaged binary vector of features; ππ β optimal radius of the
hyperspherical container
Thus, during the functioning of the ORS in the exam mode, the belonging of the feature vector is
recognized, to one of the classes from the given alphabet is determined by the decision rules (10). the
same time, the decision rules due to low computational complexity are highly efficient.
5. Results
The input training matrix in the implementation of the information-extreme machine learning
algorithm ORS with optimization of control tolerances for recognition features by procedure (3) was
formed by processing the image size 1920x1060 pixels, obtained by aerial photography of the region
from a height of 300 m (Figure 3) [17].
Figure 3: Region plan
The alphabet consisted of four recognition classes, which characterized the frames with a size of
60 Γ 60 pixels of different areas shown in Figure 2 images: class π1π β highway; class π2π β liquid
forest; class π3π β sown field; classπ4π β plowing field. Selected frames are shown in Figure 3.
a b c d
Figure 4: Image frames: a β class π1π ; b β class π2π ; c β class π3π ; d β class π4π
In order to ensure the invariance of the decision rules for the shift and rotation of objects within the
frames, the formation of the input training matrix was carried out by processing images in the polar
coordinate system. The average brightness of the pixels of each reading circle, built around the
geometric center of the frame, was calculated by the formula
1 ππ
π©π = βπ=1 ππ , (12)
ππ
where π©π is the average value of the pixels brightness included in the reading range of the π-th radius,
π = 0, π
; ππ β brightness value of the RGB component in the π-th pixel of the receptor field of the frame
image; ππ β the total number of pixels π-th reading circle; π
β radius of the reading circle.
The geometric center of the frame was determined by the formula
1+π2
ππ = πππ’ππ ( 2 ), (13)
where π is the number of pixels on the side of the square frame.
According to the averaged luminance of the reading circles calculated by formula (3), structured
vectors of recognition features of the input training matrix were formed for those shown in Figs. 3 the
region frames image.
According to the concept of IEI technology, a mandatory machine learning procedure is to optimize
the system of control tolerances for recognition features, which play the role of quantization levels in
the transformation of the input Euclidean learning matrix into a working binary learning matrix at each
step of machine learning. This raises the problem of choosing the basic class of recognition, in relation
to which is determined in the process of machine learning system of control tolerances. A working
hypothesis put forward that it is expedient to choose a recognition class as the basic one, the educational
matrix of which has the maximum variance of the brightness of the recognition features. The rationale
for this hypothesis is that the recognition class that has the largest scatter of feature brightness is the
2
closest to all classes in a given alphabet. Dispersion ππ was defined as a measure of the brightness
deviation of the π -th feature from the average value of the input training matrix brightness π©π :
1
ππ2
= (πΓπ)β1 βπΓπ Μ 2
π=1 ( π©π β π©π ) . (14)
where ππ is the brightness value of the RGB component in the π-th pixel of the image frame receptor
π
field of the recognition class ππ .
According to the results of statistical analysis of the input training matrix for a given alphabet of
recognition classes, it was found that the maximum sample variance of feature brightness was obtained
for the training matrix of recognition class π1π , relative to which the system of control tolerances was
determined. This hypothesis was experimentally confirmed by the results of information-extreme
machine learning, in which each class from a given alphabet was consistently chosen as the basic one.
Figure 5 shows a graph of the dependence of the averaged normalized criterion (8) on the parameter Ξ΄,
obtained in the process of implementing information-extreme machine learning ORS with parallel
optimization of control tolerances according to procedure (4) for the basic recognition class π1π . On the
graphs below, the working area for determining the function of criterion (7) is marked in dark color, in
which the first and second reliability exceed the errors of the first and second kind, respectively.
Figure 5: Graph of information criterion dependence on parameter of control tolerances system
The analysis of Figure 5 shows that, in the process of machine learning, the optimal value of
parameter of control tolerances system is equal to πΏ β = 43 (scale of pixelβs brightness) with a maximum
β
value of πΈ = 0,56 information criterion.
To increase the functional efficiency of machine learning, a consistent optimization of control
tolerances according to procedure (5) was implemented. In Figure 6 shows a graph of changes in the
normalized criterion in the process of sequential optimization of control tolerances for recognition
features.
Figure 6: The graph of the information criterion changes with sequential optimization of control
tolerances
Analysis of Figure 6 shows that the information optimization criterion reached a maximum value of
0.64 on the fourth run, the number of which is determined by the iterations number ratio to the features
number in the structured vectors of recognition features.
Below are graphs of the dependence of the information criterion on the radii of recognition classes
containers (9)
Π° b
c d
Figure 7: Graphs of the information criterion dependence on the radii of the recognition classes
containers: Π° β class π1π ; b β class π2π ; c β class π3π ; d β class π4π
Analysis of Figure 7 shows that the optimal values of the recognition classes containers radii are
equal to: π1β = 20 (hereinafter in code units) for class π1π ; π2β = 10 for class π2π ; π3β = 9 for class π3π
and π4β = 16 for class π4π .
Maximum values of the optimization criterion(8) that are showed on Figure 7c according to
machine learning results created on the chart plateau area. Therefore, optimum value for radii of
class π3π recognition container was considered by the minimum of ο¨ο€ coefficient which describes
overlap degree for recognition classes of given alphabet of classes:
ππ
ππΏ = β πππ, (15)
π(π₯π βπ₯π ) {π}
where π(π₯π β π₯π ) β the Hamming code distance between the averaged feature vector π₯π of the
π
recognition class ππ and corresponding feature vector π₯π of the nearest neighbor recognition class πππ .
Recognition class overlap coefficient may be considered as generalization of two basic principles
of image recognition theory: the one based on minimum distance, which demands minimization of radii
for recognition class containers and one based on maximum distances, which demands maximization
of the distances between centers of classes from given alphabet.
Figure 8 shows a digitized image (Figure 3) obtained at the stage of the exam according to the
decision rules (9).
Figure 8: The region image frames identification result
Analysis of Figure 8 shows that the identification of a highway that may be an area of interest, for
example, when searching for a vehicle, is carried out with a sufficiently high reliability. But the
constructed decision rules are not infallible. One of the ways to increase the functional efficiency of
BSR is the transition from a linear data structure to a hierarchical one with the definition for each tier
stratum of the hierarchical data structure of the basic recognition class. In addition, according to the
principle of deferred decisions of O. G. Ivakhnenko, it is necessary to increase the depth of machine
learning
6. Conclusions
Within the framework of the functional approach to modeling of cognitive processes of natural
intelligence the categorical functional model of information-extreme machine learning of ORS of
natural and infrastructural ground objects in the form of the directed graph of display of sets on each
other by machine learning operators is offered.
On the basis of the categorical model the algorithm of information-extreme machine learning with
optimization of control tolerances on recognition signs which allows to define a base class of
recognition on the greatest dispersion of the training matrix pixels brightness values is developed and
programmatically implemented. Experimentally, the results of physical modeling confirmed that the
choice of the basic recognition class, in relation to which the system of control tolerances for recognition
features is determined, directly affects the functional efficiency of machine learning.
To increase the functional efficiency of ORS it is necessary to increase the depth of machine learning
by optimizing additional parameters of the system, including the parameters of the input mathematical
description, and to increase the recognition classes alphabet the power to carry out information-extreme
machine learning hierarchical data structure.
7. References
[1] J. Iztueta, E. Lazkano, E. Martinez-Otzeta, J. Maria, S. Basilio, Visual Approaches for Handle
Recognition, Springer Tracts in Advanced Robotics 44 (2008): 313-322.
[2] H Huang, L. Lu, B.Yan, J. Chen, A new scale invariant feature detector and modified SURF
descriptor, Conference: Sixth International Conference on Natural Computation, Yantai,
Shandong, China (2010): 3734-3738.
[3] S. Kachikian, M. Emadi, Review of detector descriptorsβ on Object Tracking, International Journal
of Advanced Research in Electrical, Electronics and Instrumentation Engineering 7 (2016): 5815-
5828.
[4] V.V. Avramenko, V.M. Demianenko, Operative Recognition of Standard Signal Types, National
University Β«Zaporizhzhia PolytechnicΒ», Radio Electronics, Computer Science, Control 53
(2020):75-81. doi:10.15588/1607-3274-2020-2-8
[5] V. Moskalenko, A. Moskalenko, A. Korobov, O. Boiko, S. Martynenko, O. Borovenskyi, Model
and Training Methods of Autonomous Navigation System for Compact Drones, IEEE Second
International Conference on Data Stream Mining & Processing (DSMP), Lviv (2018): 503-508.
doi: 10.1109/DSMP.2018.8478521
[6] N. Gageik, M. Strobmeier, S. Montenegro, An autonomous UAV with an Optical Flow Sensor for
Positioning and Navigation, International Journal of Advanced Robotic Systems 10 (2013): 1 β9.
[7] A. Konert, T. Balcerzak,. Military autonomous drones (UAVs) - from fantasy to reality. Legal and
Ethical implications, Transportation Research Procedia, 59 (2021): 292β299.
doi:10.1016/j.trpro.2021.11.121
[8] V. Artale, M. Collotta, C. Milazzo, G. Pau, A. Ricciardello, Real-Time System based on a Neural
Network and PID Flight Control, Applied Mathematics and Information Sciences 10 (2016): 395-
402.
[9] M. Jafari, H. Xu, Intelligent Control for Unmanned Aerial Systems with System Uncertainties and
Disturbances Using Artificial Neural Network, Drones 2 (2018): 24-36.
[10] S. Subbotin, The neuro-fuzzy network synthesis and simplification on precedents in problems of
diagnosis and pattern recognition, Optical Memory and Neural Networks (Information Optics) 22
(2013): 97 β 103. doi: 10.3103/s1060992x13020082.
[11] V. Moskalenko, A. Moskalenko, S. Pimonenko, A. Korobov, Development of the method of
features learning and training decision rules for the prediction of violation of service level
agreement in a cloudbased environment, Eastern-European Journal of Enterprise Technologies 5
(2017): 26-33. doi: 10.15587/1729-4061.2017.110073.
[12] K. R Muller., S. Mika, G. Ratsch, K. Tsuda, B. Scholkopf, An introduction to kernelbased learning
algorithms, IEEE Transactions on Neural networks 12 (2001):181 β 202.
[13] A. S. Dovbysh, S. S. Martynenko, A. S. Kovalenko, M. M. Budnyk, Information-extreme
algorithm for recognizing current distribution maps in magnetocardiography, Journal of
Automation and Information Sciences 43 (2011): 63-70. doi:
10.1615/JAutomatInfScien.v43.i2.60.
[14] A. S. Dovbysh, M. M Budnyk, V. Yu. Piatachenko, M. I. Myronenko, Information-Extreme
Machine Learning of On-Board Vehicle Recognition System, Cybernetics and Systems Analysis
56 (2020): 534-543. doi:10.1007/s10559-020-00269-y.
[15] A. Dovbysh, I. Naumenko, M. Myronenko, T. Savchenko, Information-extreme machine learning
on-board recognition system of ground objects with the adaptation of the input mathematical
description, 3rd International Workshop on Computer Modeling and Intelligent Systems, National
University "Zaporizhzhia Polytechnic" CEUR Workshop Proceedings 2608 (2020): 913-925.
[16] I. Naumenko, M. Myronenko, T. Savchenko, Information-extreme machine training of on-board
recognition system with optimization of RGB-component digital images, Radioelectronic and
Computer Systems 98 (2021): 59-70. doi: 10.32620/reks.2021.4.05
[17] The world's most detailed globe, 2021. URL: https://www.google.com.ua/intl/en/earth/