мұндағы, х
А
, у
А
, х
В
және у
В
- жергілікті жердегі А және В тірек пункттерінің координаталары.
Содан кейін жер телімі нүктелерін анықтауға қажетті бұрыштар мәндері β
1
және β
2
мына фор-
мулалардан анықталады:
AC
AB
1
, (4)
BA
BC
2
.
Жергілікті жердегі С нүктесінің орнын анықтау үшін бастапқы бағыттан β
1
және β
2
бұрышта-
рын теодолиттің көмегімен А және В тірек пункттерінен шығарады. Осында теодолитті А және В ті-
рек пункттерінің әрқайсысына орнатып β
1
және β
2
бұрыштарын құрады да, дүрбіні бағыттайды. Осы
бағыттардың қиылысатын жақын жерінде а
1
, а
2
, в
1
, в
2
нүктелерін
екі қадамен белгілейді де, бағыттың
әрқайсысын жіппен қосады. Осы екі бағыттың қиылысатын жері жергілікті жердегі жобадағы жер
телімінің нүктесі болып саналады. Тексеру мақсатымен жер бетіндегі С нүктесіндегі γ бұрышын өл-
шейді.
Ұзындық қиылыстыру тәсілін жер телімінің жергілікті жерге шығарылатын нүктелері тірек
пункттеріне жақын орналасқан кезде, яғни тірек пункті мен нүкте арасындағы қашықтық таспаның
немесе рулетканың ұзындығынан артық болмағанда қолданылады. Жер телімінің 1 және 2 нүктелері-
нің (2,г-сурет) орындарын жер бетіне шығару мақсатымен олардың жерге орналастыру планынан
координаталарын х
1
, у
1
, және х
2
, у
2
графикалық түрде анықталады. Осыдан кейін тірек пункттерінен
жер бетіне шығарылатын жер телімінің нүктелеріне дейінгі қашықтықтарды d
1
, d
2
, d
3
және d
4
мына
формулалардан анықтайды:
2
2
1
1
1
1
1
sin
cos
Y
X
Y
Y
X
X
d
A
A
A
A
,
●
Технические науки
344
№2 2016 Вестник КазНИТУ
2
2
1
1
1
1
2
sin
cos
Y
X
Y
Y
X
X
d
B
B
B
B
, (5)
2
2
2
2
2
2
3
sin
cos
Y
X
Y
Y
X
X
d
A
A
A
A
,
2
2
2
2
2
2
4
sin
cos
Y
X
Y
Y
X
X
d
B
B
B
B
.
мұндағы, α
A-1
, α
B-1
, α
A-2
және α
B-2
– А-1, В-1, А-2 және В-2 бағыттарының дирекциондық бұрыш-
тары.
Жоғарыдағы формулада келтірілген бағыттардың дирекциондық бұрыштары α
A-1
, α
B-1
, α
A-2
және
α
B-2
мына формулалардан анықталады:
A
A
A
X
X
Y
Y
1
1
1
,
B
B
B
X
X
Y
Y
1
1
1
, (6)
A
A
A
X
X
Y
Y
2
2
2
,
B
B
B
X
X
Y
Y
2
2
2
.
Осы дайындықты орындағаннан соң жер бетінде жер телімінің 1 және 2 нүктелерінің орында-
рын анықтау үшін А және В тірек пункттерінен d
1
және d
2
, d
3
және d
4
арақашықтықтарын шығарғанда
олардың қиылысқан жерлері ізделіп отырған нүктелердің орындары болып табылады.
Жармалық тәсіл планда координаталық тор және жергілікті жерде тор төбелері бекітілген бол-
са, онда жер телімінің бұрылу нүктелері былайша анықталады (2,д-сурет). План бойынша координа-
талық тор қабырғалары бойындағы 1-1, 1-3, 1-n, 1-m, 11-n
1
, 11-m
1
, 111-2, 111-4 арақашықтықтарын
анықтайды. Жергілікті жерде анықталған осы арақашықтықтарды координаталық тор төбелерінен
квадрат қабырғаларының бойымен өлшеп салады. Содан кейін көздеу сызықтары 1-2 және n-n
1
қиы-
лысында N
1
нүктесін, 1-2 және m-m
1
қиылысында M
1
нүктесін, 3-4 және n-n
1
қиылысында N нүктесін,
3-4 және m-m
1
қиылысында M нүктесін табады.
Жергілікті нысандардан бөлу тәсілі ішінара құрылыстанған аумақ ішінде жер телімі орналас-
қанда (2,е-сурет), оның бұрылу бұрыштары жергілікті жердегі нысандардан жоғарыдағы келтірілген
тікбұрышты және полярлық координаталар, бұрыштық және ұзындық қиылыстырулар немесе осы
тәсілдердің қисындасуымен анықталады. Осында барлық бастапқы мәліметтер планнан графикалық
түрде анықталады, себебі жер телімі ғимараттар мен нысандар арасында болған кезде бөлулік жұ-
мыстарды жүргізуде жоғары дәлдіктің қажеттігі тумайды.
ӘДЕБИЕТТЕР
[1] Қазақстан Республикасының Жер кодексі. –Алматы: Жеті жарғы, 2003. -256 б.
[2] Поклад Г.Г., Гриднев С.П. Геодезия. –М.: Академический Проект, 2007.-592 c.
REFERENCES
[1] Kazakhstan Respublikasynyn Zher kodeksi. – Almaty: Zheti zhargy, 2003. – 256 b.
[2] Poklad G.G., Gridnev S.P. Geodeziya. – M: Akademicheskiiy Proekt, 2007. – 592 s.
●
Техникалық ғылымдар
ҚазҰТЗУ хабаршысы №2 2016
345
Калыбеков Т., Абен А.С.
Способы перенесения в натуру проектных границ земельного участка
Резюме. В статье приведены способы перенесения в натуру при землеустроительных работах проектных
границ земельных участков. Приведены формулы для определения координат проектных точек земельного
участка при землеустройтельных работах.
Ключевые слова: земельный участок, граница, кадастр, землеустройство, перенесение в натуру.
Kalybekov T., Aben A.S.
Methods of transference in nature of project borders of lot land
Summary. Ways of transferring to nature are given in article during the land management works of design
borders of the land plots. The formulas for determining the coordinates of the design point of land in the land surveying
work.
Key words: land plot, border, inventory, land management, transferring to nature.
UDC 519.68
N. Seilova
1
L. Tereykovskaya
2
, A. Наджи
3
(
1
Kazakh national research technical university named after Satpayev K.I.,
Almaty, Kazakhstan. seilova_na@mail.ru
2
Kyiv National University of Construction and Architecture, Kyiv, Ukraine, terejkowski@ukr.net
3
National Aviation University, Kyiv, Ukraine, abdonagi@hotmail.com)
CONCEPTUAL MODEL TO ENSURE THE EFFICIENCY OF NEURAL NETWORK
RECOGNITION OF PHONEMES IN DISTANCE LEARNING
Abstract. The article is devoted to the problem of improving the effectiveness of distance learning through the
introduction of interactive training materials, based on the media to recognize the voice signals. It is shown that the pro-
spects for the development of such tools related to the creation of neural networks, which are designed to recognize
phonemes and should be adapted for use in distance learning. An outline promising research paths towards development
of effective means of neural network using a conceptual model. The conceptual model to ensure the efficiency of neural
network recognition of phonemes in distance learning. Unlike the existing conceptual model takes into account the fact
that the expected terms of introduction of neural network means characterized by variability limits on term develop-
ment, attracting labor resources acoustic parameters of the voice signal computing resources web server system of dis-
tance education and restrictions on the use of educational databases neural network models. In the process of creating a
model of the factors affecting the performance of neural network recognition of phonemes in distance learning and
identified a number of indicators to measure the efficiency of the process. Using the model allows us to go to develop
the methodological framework creating effective means of neural network recognition.
Key words: voice recognition signal, distance learning, neural network, phoneme, conceptual model.
Formulation of the problem. Today it is believed that one of the important ways to improve the ef-
fectiveness of distance learning is the use of interactive learning tools based on mass recognition of voice
signals [6]. Note that commonly recognize voice signals means companies such as Google and Microsoft are
based on neural network models. Together with the findings [1-3], indicating promising application of neural
network models and the means to recognize the voice signals of distance learning. At the same time the most
famous of distance learning neural network means of identification available. Therefore presents scientific
and practical interest in the sector create effective means of neural network to recognize the voice signals
adapted for use in distance learning.
Analysis of recent research and publications. Accordingly, [6], the voice interaction in the operation
of distance learning should be used during lectures, seminars, consultations, laboratory and practical classes.
It is a natural addition to this list is the use of voice interaction for user authentication. Note that during a
voice interaction will understand the interaction between the components of distance learning, which is
based on the recognition of voice signals. In many cases, this recognition should be conducted in automatic
mode, for example, computer testing, the student must provide voice answer the question. This recognition
system must simultaneously solve two problems: to identify the correct answer and perform user authentica-
tion.
●
Технические науки
346
№2 2016 Вестник КазНИТУ
The solution to both problems of voice recognition based on user information. Note that in general
such recognition is consistent with the dual tasks:
- Creating formal description of voice.
- Conduct semantic analysis received formal description.
Quite often during formal descriptions of voice understand its text representation. This modern theo-
retical developments in the field of semantic analysis of text information do not allow you to create highly
reliable tools, and in many cases the student response and person can be determined on the basis of detection
(absence) it more specific words. Thus, concerning information technology distance learning, recognition of
voice is to find it and keyword recognition speaker. Consider the problem of search keywords. Accordingly,
[9], this search algorithm generally consists of the following stages:
Digitization reference and experimental audio signal.
Filtering noise.
With bold signal words.
Digitized signal processing to reduce the amount of input recognition system.
Additional filtration spectrum.
Compression range to incorporate features of human perception of sound and reduce the number
of input parameters recognition system.
Comparison of experimental and reference signals.
Under [3-5, 9], the most difficult stage of the search is to implement recognition procedure, which re-
sults in standard definition that most closely reflects incoming (unknown) voice signal. Difficulty recogni-
tion procedure, primarily due to the fact that the voice signal is characterized by non-linear change of pace
broadcasting of words and different duration of pauses at the beginning and end of words. Therefore, in most
cases directly compare the unknown voice signal with standard word impossible. Typically, the recognition
procedure is divided into several stages. Incoming voice signal is first divided into elements - phonemes
(phonemes phase), allophones, dyfony, Trifon, warehouses, etc. For these elements are most similar stand-
ards and standards already using elements are most similar patterns of individual words. Section voice signal
into separate elements efficiently performed by analyzing energy components. Methods recognition of indi-
vidual words based on standards elements are also considered sufficiently proven and reliable [9,10]. At the
same time the problem of the individual elements of the standards is far from solved. Results [1,5] suggest
about the prospects for use as individual elements phonemes, because of their relative malochyselnistyu
compared to the number of warehouses, allophones, and dyfoniv triphones. In accordance with the conclu-
sions [2] to recognize phonemes are useful neural network means. In this case the expected introduction of
conditions characterized by means of neural variability limits on term development, attracting labor re-
sources acoustic parameters of the voice signals and computing resources on a web server of distance learn-
ing. It should be noted restrictions on the use of databases examples recordings necessary for the study of
neural network models, which greatly affects recognition accuracy means that are created based on them.
However, analysis of the literature [4,8,10] indicates insufficient adaptation of modern means of neural net-
work recognition to the voice signal variability appointed in terms of implementation of distance learning. In
turn, the lack of adaptation could affect the effectiveness of such measures. The purpose of this paper is to
develop a conceptual model to ensure the efficiency of neural network recognition of phonemes in distance
learning.
Development of a conceptual model. In general, the conceptual model is a model subject area, consist-
ing of a list of related concepts that are used to describe this area along with the properties and characteristics
of these concepts classification by type situations featured in the art, and the laws it flow processes [6,7].
Conceptual model is a reflection of the concept, the notion of which way to understand certain opinions, the
interpretation of certain phenomena, the basic point of view, leading to the idea of systematic coverage. We
note that the development of a conceptual model is a common starting point of the methodological frame-
work, which is a system of principles and methods of organization and construction of theoretical and practi-
cal work and theory of the system. Because the bottom line thesis involves the creation of software and
hardware for recognition of phonemes, to determine the effectiveness of the neural network recognition of
phonemes in the voice alarm system provides distance learning to use the definition of the field of computer
and software engineering. Under international standards that area efficiency - the set of attributes that define
the relationship level execution systems, use of resources (tools, equipment, materials - paper for printers,
●
Техникалық ғылымдар
ҚазҰТЗУ хабаршысы №2 2016
347
etc.) And services performed by staff and other support staff. By the performance characteristics of a soft-
ware system include:
efficiency - an attribute that indicates the response time of processing and fulfillment functions;
specific resource consumption - an attribute that defines the amount and duration of resources
used in the performance of specified functions;
consistency - an attribute that indicates compliance with this attribute set standards, rules and
regulations.
In accordance with the definitions, the first stage of creating conceptual model was conducted harmo-
nize terminology used in the application of neural networks for voice recognition signal. Harmonization of
positions held display the state of the practice and supports problem solving thesis. As a result, the following
terms:
- Voice is - a complex acoustic signal, the source of which is the human voice. In the context of this
thesis synonymous voice signal is a voice signal, although in general terms between the data there are some
differences [6].
- Phoneme similar element - highlighted in the voice signal fragment, whose parameters correspond to
separate phonemes.
- Phoneme - the minimal structural and functional unit of speech sound, which is used for the identifi-
cation of differences and meaningful units of language.
- Neural network - a network of artificial neurons connected by synaptic (suspended) bonds [7].
- Neural network model - a model of a neural network, characterized by learning, signal propagation
method, structure type connections and artificial neurons.
These options and their combinations determine the type of neural network models. Synonymous with
the type of neural network model is a neural network architecture. Derived from the term neural network
model is a neural network techniques, neural network system and neural network means that the methods,
systems and tools that are based on the use of neural networks. As generally understood by the term means
instrument (object, device, collection of devices), the concept of neural network is a collective means for
neural models and neural systems used for recognition of phonemes in the voice signal of distance learning.
This hardware and software implementation of such devices will be called instrumental neural network tool.
Also determined that the problem regarding this dissertation conceptual model is intended to formalize the
causal relationships that are inherent in the process of recognition of phonemes in the voice signal is deter-
mined by the need to improve the efficiency of distance learning. In addition, the conceptual model should
take into account:
- conditions for the functioning of neural network by means of recognition of phonemes nature of the
interaction between components and remote training, as well as between its different parts;
- the need to implement effective application of neural network models for recognition of phonemes
and direction of its improvement;
- neural network means more control and determination of its controlled variables.
Therefore, the next phase of construction of a conceptual model was presented contextual process dia-
gram of neural network recognition of phonemes in the voice signal, reflecting its basic function and interac-
tion with the environment (see. Fig. 1).
Достарыңызбен бөлісу: |