Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities
Video surveillance has become an important constituent part of the integrated security system of nuclear power plants. Due to this, the integrity and authenticity of the video transmitted by the surveillance camera are extremely important, and so is the possibility to identify violations of these ca...
Saved in:
| Published in: | Ядерна та радіаційна безпека |
|---|---|
| Date: | 2016 |
| Main Authors: | , , , |
| Format: | Article |
| Language: | English |
| Published: |
Державне підприємство "Державний науково-технічний центр з ядерної та радіаційної безпеки" Держатомрегулювання України та НАН України
2016
|
| Online Access: | https://nasplib.isofts.kiev.ua/handle/123456789/129810 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Journal Title: | Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| Cite this: | Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities / I. Bobok, A. Kobozeva, M. Maksymov, O. Maksymova // Ядерна та радіаційна безпека. — 2016. — № 2. — С. 68-72. — Бібліогр.: 11 назв. — англ. |
Institution
Digital Library of Periodicals of National Academy of Sciences of Ukraine| id |
nasplib_isofts_kiev_ua-123456789-129810 |
|---|---|
| record_format |
dspace |
| spelling |
Bobok, I. Kobozeva, A. Maksymov, M. Maksymova, O. 2018-01-29T16:27:04Z 2018-01-29T16:27:04Z 2016 Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities / I. Bobok, A. Kobozeva, M. Maksymov, O. Maksymova // Ядерна та радіаційна безпека. — 2016. — № 2. — С. 68-72. — Бібліогр.: 11 назв. — англ. 2073-6231 https://nasplib.isofts.kiev.ua/handle/123456789/129810 004.056.5 Video surveillance has become an important constituent part of the integrated security system of nuclear power plants. Due to this, the integrity and authenticity of the video transmitted by the surveillance camera are extremely important, and so is the possibility to identify violations of these categories of information in real time. The authors propose a new method to detect one way of violating the integrity of the video sequence — “camera loop” hijacking. The proposed method based on identifying the perturbation in matrix of the current frame of the original video sequence in transition to the next frame ensures the efficiency of the method that is not dependent on the specifics and characteristics of the analyzed video sequence. The high efficiency of the method is confirmed by results of a computational experiment, under which type I and type II errors were not found. The practical value of the proposed method is the possibility of its work in real time because it is a polynomial of degree 1, as well as the simplicity and multiplatform implementation. За останні роки відеоспостереження стало невід’ємною складовою частиною комплексної системи безпеки атомних електростанцій. Тому надзвичайно важливими є цілісність і автентичність відеоінформації, переданої камерою відеоспостереження, а також можливість встановлення порушень згаданих категорій інформації в режимі реального часу. У роботі пропонується метод виявлення одного з порушень цілісності відеопослідовності — заставки. Принцип, покладений в основу методу (виявлення збурень матриці поточного кадру оригінальної відеопослідовності під час переходу до подальшого кадру), дає змогу забезпечити незалежність ефективності методу від специфіки отримання та характеристик аналізованої відеопослідовності. Висока ефективність методу підтверджена результатами обчислювального експерименту, в умовах якого помилки першого і другого роду не виявлено. Практична цінність методу полягає в можливості організації його роботи в режимі реального часу, оскільки він є поліноміальним ступеня 1, а також у простоті й мультіплатформенності реалізації. За последние годы видеонаблюдение стало неотъемлемой составной частью комплексной системы безопасности атомных электростанций. В силу этого чрезвычайно важными являются целостность и аутентичность видеоинформации, передаваемой камерой видеонаблюдения, а также возможность установления нарушений упомянутых категорий информации в режиме реального времени. В работе предлагается метод выявления одного из способов нарушения целостности видеопоследовательности — заставки. Принцип, положенный в основу предложенного метода, состоящий в выявлении возмущений матрицы текущего кадра оригинальной видеопоследовательности при переходе к последующему кадру, позволяет обеспечить независимость эффективности метода от специфики получения и характеристик анализируемой видеопоследовательности. Высокая эффективность метода подтверждена результатами вычислительного эксперимента, в условиях которого ошибки первого и второго рода обнаружены не были. Практическая ценность предложенного метода заключается в возможности организации его работы в режиме реального времени, поскольку он является полиномиальным степени 1, а также в простоте и мультиплатформенности реализации. en Державне підприємство "Державний науково-технічний центр з ядерної та радіаційної безпеки" Держатомрегулювання України та НАН України Ядерна та радіаційна безпека Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities Перевірка цілісності записів камер відеоспостереження в режимі реального часу на об’єктах атомної енергетики Проверка целостности записей камер видеонаблюдения в режиме реального времени на объектах атомной энергетики Article published earlier |
| institution |
Digital Library of Periodicals of National Academy of Sciences of Ukraine |
| collection |
DSpace DC |
| title |
Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities |
| spellingShingle |
Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities Bobok, I. Kobozeva, A. Maksymov, M. Maksymova, O. |
| title_short |
Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities |
| title_full |
Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities |
| title_fullStr |
Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities |
| title_full_unstemmed |
Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities |
| title_sort |
checking the integrity of cctv footage in real time at nuclear facilities |
| author |
Bobok, I. Kobozeva, A. Maksymov, M. Maksymova, O. |
| author_facet |
Bobok, I. Kobozeva, A. Maksymov, M. Maksymova, O. |
| publishDate |
2016 |
| language |
English |
| container_title |
Ядерна та радіаційна безпека |
| publisher |
Державне підприємство "Державний науково-технічний центр з ядерної та радіаційної безпеки" Держатомрегулювання України та НАН України |
| format |
Article |
| title_alt |
Перевірка цілісності записів камер відеоспостереження в режимі реального часу на об’єктах атомної енергетики Проверка целостности записей камер видеонаблюдения в режиме реального времени на объектах атомной энергетики |
| description |
Video surveillance has become an important constituent part of the integrated security system of nuclear power plants. Due to this, the integrity and authenticity of the video transmitted by the surveillance camera are extremely important, and so is the possibility to identify violations of these categories of information in real time. The authors propose a new method to detect one way of violating the integrity of the video sequence — “camera loop” hijacking. The proposed method based on identifying the perturbation in matrix of the current frame of the original video sequence in transition to the next frame ensures the efficiency of the method that is not dependent on the specifics and characteristics of the analyzed video sequence. The high efficiency of the method is confirmed by results of a computational experiment, under which type I and type II errors were not found. The practical value of the proposed method is the possibility of its work in real time because it is a polynomial of degree 1, as well as the simplicity and multiplatform implementation.
За останні роки відеоспостереження стало невід’ємною складовою частиною комплексної системи безпеки атомних електростанцій. Тому надзвичайно важливими є цілісність і автентичність відеоінформації, переданої камерою відеоспостереження, а також можливість встановлення порушень згаданих категорій інформації в режимі реального часу. У роботі пропонується метод виявлення одного з порушень цілісності відеопослідовності — заставки. Принцип, покладений в основу методу (виявлення збурень матриці поточного кадру оригінальної відеопослідовності під час переходу до подальшого кадру), дає змогу забезпечити незалежність ефективності методу від специфіки отримання та характеристик аналізованої відеопослідовності. Висока ефективність методу підтверджена результатами обчислювального експерименту, в умовах якого помилки першого і другого роду не виявлено. Практична цінність методу полягає в можливості організації його роботи в режимі реального часу, оскільки він є поліноміальним ступеня 1, а також у простоті й мультіплатформенності реалізації.
За последние годы видеонаблюдение стало неотъемлемой составной частью комплексной системы безопасности атомных электростанций. В силу этого чрезвычайно важными являются целостность и аутентичность видеоинформации, передаваемой камерой видеонаблюдения, а также возможность установления нарушений упомянутых категорий информации в режиме реального времени. В работе предлагается метод выявления одного из способов нарушения целостности видеопоследовательности — заставки. Принцип, положенный в основу предложенного метода, состоящий в выявлении возмущений матрицы текущего кадра оригинальной видеопоследовательности при переходе к последующему кадру, позволяет обеспечить независимость эффективности метода от специфики получения и характеристик анализируемой видеопоследовательности. Высокая эффективность метода подтверждена результатами вычислительного эксперимента, в условиях которого ошибки первого и второго рода обнаружены не были. Практическая ценность предложенного метода заключается в возможности организации его работы в режиме реального времени, поскольку он является полиномиальным степени 1, а также в простоте и мультиплатформенности реализации.
|
| issn |
2073-6231 |
| url |
https://nasplib.isofts.kiev.ua/handle/123456789/129810 |
| citation_txt |
Checking the Integrity of CCTV Footage in Real Time at Nuclear Facilities / I. Bobok, A. Kobozeva, M. Maksymov, O. Maksymova // Ядерна та радіаційна безпека. — 2016. — № 2. — С. 68-72. — Бібліогр.: 11 назв. — англ. |
| work_keys_str_mv |
AT boboki checkingtheintegrityofcctvfootageinrealtimeatnuclearfacilities AT kobozevaa checkingtheintegrityofcctvfootageinrealtimeatnuclearfacilities AT maksymovm checkingtheintegrityofcctvfootageinrealtimeatnuclearfacilities AT maksymovao checkingtheintegrityofcctvfootageinrealtimeatnuclearfacilities AT boboki perevírkacílísnostízapisívkamervídeosposterežennâvrežimírealʹnogočasunaobêktahatomnoíenergetiki AT kobozevaa perevírkacílísnostízapisívkamervídeosposterežennâvrežimírealʹnogočasunaobêktahatomnoíenergetiki AT maksymovm perevírkacílísnostízapisívkamervídeosposterežennâvrežimírealʹnogočasunaobêktahatomnoíenergetiki AT maksymovao perevírkacílísnostízapisívkamervídeosposterežennâvrežimírealʹnogočasunaobêktahatomnoíenergetiki AT boboki proverkacelostnostizapiseikamervideonablûdeniâvrežimerealʹnogovremeninaobʺektahatomnoiénergetiki AT kobozevaa proverkacelostnostizapiseikamervideonablûdeniâvrežimerealʹnogovremeninaobʺektahatomnoiénergetiki AT maksymovm proverkacelostnostizapiseikamervideonablûdeniâvrežimerealʹnogovremeninaobʺektahatomnoiénergetiki AT maksymovao proverkacelostnostizapiseikamervideonablûdeniâvrežimerealʹnogovremeninaobʺektahatomnoiénergetiki |
| first_indexed |
2025-11-26T21:28:55Z |
| last_indexed |
2025-11-26T21:28:55Z |
| _version_ |
1850776787024347136 |
| fulltext |
68 ISSN 2073-6231. Ядерна та радіаційна безпека 2(70).2016
M
odern digital technologies, which are applied
in the development of automated control systems
of technological processes, computer networks,
security systems, etc., play an important
role in the operation of nuclear power plants.
These technologies transmit information from the station level
to the power grid management level and back, implement
the operational management of the current regime by operators,
track changes in the state of an object in real time, etc. It should
be noted that the use of the above-mentioned technologies, many
of which are vulnerable to cyber attacks, at installations with
limited access requires special attention to the issue of integrated
security [1, 2].
This paper considers one element of the integrated security
system — CCTV system.
Early Basic Research Efforts and Publications. Today,
CCTV has become an essential part of the integrated security
system at limited access facilities (banks, administrative
institutions, scientific laboratories, nuclear power plants),
because the modern equipment allows not only monitoring
and control of various components of the object in real time
and video recording, but also program response from the entire
security system upon alarm signals. In view of this, camera-
generated video sequence may be the subject of falsification by
the attackers, whose aim is the unauthorized entry to a secure
object.
The falsifications of the video sequences from surveillance
cameras can occur in various ways [3]. First of all, it depends
on where the camera (indoor/outdoor) is placed and where its
stationary position is. So, if the location of a stationary camera
is indoors, access to which is limited, an attacker can involve
hijacking video streams and seamlessly replace them with static
image which repeats over and over again; it is called “camera
loop” [4].
The Aim of Research is to develop a method for detection
of the “camera loop” hijacking in video sequences from
surveillance cameras in real time.
Description of Research. In view of the fact that the use
of the “camera loop” hijacking is limited in time (often small)
interval t, it is extremely difficult to visually notice the result
because the dynamics of scene changes in the premises
where movement is very small. However, in a real room
this dynamics should take place, which is associated with
(albeit minor) heat and mass transfer, brightness, humidity,
etc., which leads to movement of air masses. Modern video
cameras are very sensitive to the most minor changes of scene,
invisible to the human eye. These changes can be registered
as perturbation of brightness matrix elements of consecutive
frames of video sequences.
To confirm the hypothesis, numerical experiments were
conducted. During the experiments, various cameras were
installed in the inaccessible premises and made 24th-shooting
of all scenes. Moreover, the presence of natural and artificial
light sources in the room changed with different (stationary)
position of the camera geometrically and toward the light
source. The arbitrary video sequence is a sequence of frames,
each of which can be seen as a stationary digital image.
The formal representation of such images is one, three, four
two-dimensional matrices depending on color or monochrome
image, storage schema, etc. [3]. The changes of scene in terms
of developments can be seen as changes of element values
of matrices, corresponding to successive frames.
During the experiment, the first 100 frames from each
analyzed video sequence were thrown out, and then sequence
of 200 frames was processed. For each pair of consecutive
frames, n×m-matrices of which are denoted as F and F with
УДК 004.056.5
I. Bobok, A. Kobozeva, M. Maksymov,
O. Maksymova
Odessa National Polytechnic University, Odessa, Ukraine
Checking the integrity
of CCTV footage in real time
at nuclear facilities
Video surveillance has become an important constituent part of the in-
tegrated security system of nuclear power plants. Due to this, the integrity
and authenticity of the video transmitted by the surveillance camera are
extremely important, and so is the possibility to identify violations of these
categories of information in real time.
The authors propose a new method to detect one way of violating the in-
tegrity of the video sequence — “camera loop” hijacking. The proposed
method based on identifying the perturbation in matrix of the current frame
of the original video sequence in transition to the next frame ensures the ef-
ficiency of the method that is not dependent on the specifics and character-
istics of the analyzed video sequence. The high efficiency of the method is
confirmed by results of a computational experiment, under which type I and
type II errors were not found. The practical value of the proposed method is
the possibility of its work in real time because it is a polynomial of degree 1,
as well as the simplicity and multiplatform implementation.
K e y w o r d s: integrated security system, video surveillance, video se-
quence, camera loop.
І. І. Бобок, А. А. Кобозєва, М. В. Максимов, О. Б. Максимова
Перевірка цілісності записів камер відеоспостере-
ження в режимі реального часу на об’єктах атомної
енергетики
За останні роки відеоспостереження стало невід’ємною складовою
частиною комплексної системи безпеки атомних електростанцій. Тому
надзвичайно важливими є цілісність і автентичність відеоінформації,
переданої камерою відеоспостереження, а також можливість встанов-
лення порушень згаданих категорій інформації в режимі реального часу.
У роботі пропонується метод виявлення одного з порушень ціліс-
ності відеопослідовності — заставки. Принцип, покладений в осно-
ву методу (виявлення збурень матриці поточного кадру оригінальної
відео послідовності під час переходу до подальшого кадру), дає змогу
забезпечити незалежність ефективності методу від специфіки отри-
мання та характеристик аналізованої відеопослідовності. Висока ефек-
тивність методу підтверджена результатами обчислювального експе-
рименту, в умовах якого помилки першого і другого роду не виявлено.
Практична цінність методу полягає в можливості організації його ро-
боти в режимі реального часу, оскільки він є поліноміальним ступеня 1,
а також у простоті й мультіплатформенності реалізації.
К л ю ч о в і с л о в а: комплексна система безпеки, відеоспостере-
ження, відеопослідовність, порушення цілісності, заставка.
© I. Bobok, A. Kobozeva, M. Maksymov, O. Maksymova, 2016
ISSN 2073-6231. Ядерна та радіаційна безпека 2(70).2016 69
Checking the integrity of CCTV footage in real time at nuclear facilities
elements ,ij ijf f , 1, , 1,i n j m= = , respectively, we determine
the n×m-matrix of absolute difference ( )abs= −R F F with
elements
,ij ij ijr f f= − 1, , 1,i n j m= = .
We founded that for each pair of consecutive frames from
each of the original video sequences, there are elements
of the matrix R different from 0, i.e. for each pair of consecutive
frames:
≠F F .
Moreover, such elements constitute from 8 to 21 % of the total.
Detailed experimental results for some of the examined video
sequences are presented in Table 1. The results are given
for the red color component of the matrix representation
of the frame (RGB system); those for the other two color
components are comparable to the presented results.
On average, for each pair of consecutive frames, brightness
of nine percent pixels is changed by more than one scale.
A typical pattern is shown in Fig. 1.
The chart corresponds to the original video sequence that
fully confirms the proposed hypothesis. For the “camera
loop”, there are no changes from frame to frame, which would
allow distinguishing it from the original video and identifying
unauthorized actions of the attacker.
Taking into account the potential large size of analyzed
frames and the need for fundamental possibility of analyzing
in real-time, it is possible to analyze not entire frame, but only
part of it. To do this, we must set the allowable size of the part
of the frame, the analysis of which would be informative in terms
of detecting “camera loop” hijacking, i.e. the portion sizes,
which will obligatory contain disturbed pixels (for the frame
sequence number of such pixels may be less than 5 % (Table 1)
and their arrangement within a frame in the general case is
unknown).
As a result of computing experiment, it was determined
that for video sequence corresponding to scene with natural
light source it is sufficient to consider the frame’s sub-domains
of size 100 × 100 pixels (such sub-domains necessarily contain
(single) disturbed pixels) (Fig. 2, b), and without natural light
source — 32 × 32 pixels (Fig. 2, a).
To reduce the analysis time of the video sequence (to allow
the execution in real time) we propose to conduct a comparison
of pairs of the non-consecutive frames, but frames taken with
certain time period T. It is necessary that
T < t. (1)
Table 1. Results of the computational experiment for
original video sequences made by stationary camera
VS
Arithmetical mean
of the number
of non-zero
elements of matrix R
by frames of video
sequence, %
Arithmetical
mean for maximal
value of elements
of matrix R
by frames of video
sequence
Arithmetical mean
of the number
of non-zero
elements of matrix R
greater than one,
%
V1 81 31 10
V2 82 33 11
V3 81 27 10
V4 84 27 9
V5 90 33 5
V6 90 24 5
V7 90 25 4
V8 89 32 5
V9 87 36 8
V10 83 37 10
Fig. 1. The typical form of histogram of the values of the elements of matrix R for two consecutive frames of the original video sequence:
a — without natural light source; b — with natural light source
а b
70 ISSN 2073-6231. Ядерна та радіаційна безпека 2(70).2016
I. Bobok, A. Kobozeva, M. Maksymov, O. Maksymova
Using the non-consecutive frames will reduce analysis time
due to:
the reduction in the number of analyzed pairs of frames;
the fact that differences in non-consecutive frames of the original
video sequences will be in more pixels than in consecutive frames; it
will also reduce the sizes of the analyzed subdomains of frames.
In view of the presented results, algorithm of detecting
the presence of the “camera loop” in the video sequence from
surveillance camera in real time is as follows.
Let’s V is analyzed video sequence, where V (1), V (2), …, V (k)
are frames; r is frame rate; T is a period between two analyzed
frames and is measured in seconds.
Step 1 (Initialization). p = p0 (p0 — number of first analyzed
frame from V ).
Step 2. For V (p) and V (p+rT ) we extract the analyzed
subdomains of size l × l: F(l) and
( )l
F in an arbitrary color
component (RGB system) (or brightness matrix (YUV [5])).
Step 3. Forming the l × l-matrix abs
( )( ) ll = − R F F .
Step 4. For matrix R elements:
If rij = 0 for , 1,i j l∀ = ,
then “Camera Loop” detected, EXIT;
else
if 2p rT k+ ≤
then p = p + rT, go to Step 2
else EXIT.
Remark. For processing the pair of frames of the video
sequence, the number of operations does not depend on the size
of the frame’s matrix, the total number of frames, i.e. it is
a constant. In view of this, the maximum computational
complexity of the algorithm in the processing of frames V(1),
V(2), …, V(k) (in the case of processing two successive frames, i.e.
rT = 1) would be defined as
( )O k , (2)
which makes it possible to use it in real time.
Discussion. Underlying the principle of the proposed method
based on identifying the perturbation of matrix of the frame
of the original video sequence in the transition to the next
frame, ensures the independence of the efficiency of the method
on the specifics of obtaining and characteristics of the analyzed
video sequence.
To test the effectiveness of the developed method
and algorithmic implementations for different values
of the parameters numerical experiments were conducted
in MATLAB 7.4.0.287 (License number: 21808. Platform:
All. License option: Group Term: Perpetual. Use: Classroom)
on a 2.7 GHz Intel Core i5 processor. There were 300 original
video sequences (MPEG-4 codec, MOV container), each
of which was from 90 to 900 seconds recorded by multiple
cameras, namely Olympus SP-820 — 1/2,3” CMOS, 14MP;
Nikon COOLPIX P100 — 1/2,3” CMOS, 10,3MP; Canon
PowerShot A520 — 1/2,5” CCD, 4MP. It should also be
pointed out that characteristics of cameras used in experiments
similar to cameras used in production facilities including at
nuclear facilities, for example, Rolls-Royce CDS 5000 [6], or
ECA Group DTR 65 HRC [7]. Shooting was made at different
times of the day, in presence/absence of natural/artificial light
sources (Table 2 — here, the last column shows the designation
of the video sequence belonging to this category, which is used
below); herewith frames obtained by cameras are represented
as images with different resolution (from 320×240 to 1920×1080),
contrast, brightness, with/without fine details.
The proposed method is actually a binary classifier
therefore the effectiveness of the algorithmic implementation
of the method was evaluated using the Type I (“camera loop”
is not detected) and Type II (“camera loop” is false detected)
errors.
Type I and Type II errors were defined by a standard way [8],
namely, the frequency of occurrence of the relevant events
in the experiments, taking into account that with an increase
in number of trials (in our case — the number of analyzed
video sequences) the frequency of the event loses its random
nature, tends to stabilize, and approach to some average
Fig. 2. The typical form of histogram of the values of the elements of matrix R for subdomains of two consecutive frames of the original video sequence:
a — without natural light source size of subdomain is 32×32 pixels; b — with natural light source (size of subdomain is 100×100 pixels)
а b
ISSN 2073-6231. Ядерна та радіаційна безпека 2(70).2016 71
Checking the integrity of CCTV footage in real time at nuclear facilities
constant value (in accordance with Bernoulli’s theorem), which
is the probability of the event. It is only necessary to note that
this “approach” has a feature: with increasing number of trials n
the frequency of event un converges to its probability p, i.e. for
any ε > 0 with increasing n the probability of the following
inequality
nu p− < ε
tends to 1 [9]. Thus, with increasing number of trials n
the frequency approaches to the probability without absolute
certainty, but with the probability that with large number
of trials can be considered as the practical certainty [9]; that
was used for estimating of Type I and Type II errors.
Table 2. The distribution of 300 original video sequences used
during the numerical experiment into different categories
Number of video sequences that
made in the presence of natu-
ral light (without artificial light
source)
Time 600 — 1200 25 V11
Time 1200 — 1700 25 V12
Time 1700 — 2400 25 V13
Number of video sequences
that made in the presence
of natural light (with artificial
light source)
Time 600 — 1200 25 V21
Time 1200 — 1700 25 V22
Time 1700 — 2400 25 V23
Number of video sequences
that made without natural light
source (with artificial light
source)
150
V31
The theoretical possibility of Type I errors can be only
in one case — if the condition (1) is false; in any other case
Type I errors are impossible. If the preliminary assessment
of the time t of the attacker’s possible actions is not available,
the best in order to avoid a possible “camera loop” hijacking
skip during the video sequence analysis is to consider the pairs
of consecutive frames, i.e.
1
T
r
= guarantees the absence
of errors of Type I errors.
Type II errors [8] are fundamentally possible in the case
where the sizes l × l of the analyzed subdomain of the frame
would be insufficient to contain guaranteed perturbed (relative
to the previous frame pixels) of the subsequent analyzed frame
of the original video sequence. Results of computational
experiment where 1
T
r
= , are given in Table 3; the conclusion
of the “camera loop” in video sequences was made as soon
as was found at least one pair of frames for which l×l-matrix
R = 0.
The presented results in practice confirm the above
theoretical conclusion that the effectiveness of this method does
not depend on the specific conditions in which the received
video sequence, the specifics of frames (resolution, contrast,
brightness, presence/absence and size of the background
areas, presence/absence of fine details) — the results for all
categories are congruent. In the algorithmic implementation
of the method, it makes sense to consider l×l-subdomains for
l ≥ 100.
Table 3. Type II errors in the analysis of video
sequences using the developed method, %
l
VS
8 16 32 100 120
V11 100 87 14 2 0
V12 97 87 17 0 0
V13 98 86 12 0 0
V21 100 84 15 0 0
V22 99 72 21 2 0
V23 99 88 11 1 0
V31 98 90 2 0 0
Conclusions. Securing objects with limited access, which are
the modern nuclear power plants, remained unresolved task until
now and requiring solutions for the integrated systems approach.
Certainly, the special attention is given to the important
issues relating to nuclear security and preventing interference
from the outside, cooling the reactor and prevent the escape
of radioactive substances outside the pressurized zone [10], while
the cause of such worst-case situations may be the unauthorized
actions of attackers who had access to the limited access
premises and networks. Due to this, the integrity and authenticity
of the video transmitted by the CCTV cameras are extremely
important, as well as the possibility of establishing violations
of these categories of information in real time.
In this article the authors presented a new method for
detection of the “camera loop” hijacking in video sequences
from surveillance cameras in real time. The high efficiency
of the developed method is confirmed by the results
of computational experiments — during the analysis of video
sequences by comparing the l × l-subdomains of the pairs
of consecutive frames with l > 100, Type I and Type II errors
were not found.
The practical value of the proposed method is as follows:
• the possibility of organizing its work in real time,
as in accordance with (2) it is a polynomial of degree 1, even
when processing successive frames of a video sequence;
• simplicity and multiplatform implementation.
The signal from the CCTV cameras considered by
the way of attacks on the integrity is one of the most easily
implemented variations of this violation. This method is not
used by the attackers if camera’s installation is not stationary,
or the camera is outside the premise with limited access. In this
case, it will be used “insert”-method [11], when the original
signal V from the camera at the time t will be replaced by another
video signal tV , and then the broadcast of original signal
will be resumed. There are several options for generating tV ,
but the most interesting, difficult to recognize and most likely
(given to the interest in challenging the process of establishing
the real signal substitution) is a choice when tV video sequence
showing the scene similar to V and made by the same camera for
which substitution is performed. Here, the problem of detecting
the substitution is reduced to determine of “place of insert”,
because all the technical parameters of the video sequence tV
are similar to the parameters of the original video V. The results
of this task solution in real-time are currently being prepared
for publication by the authors.
72 ISSN 2073-6231. Ядерна та радіаційна безпека 2(70).2016
I. Bobok, A. Kobozeva, M. Maksymov, O. Maksymova
References
1. Song, J.-G., Lee, J.-W., Lee, C.-K. (2012), “A Cyber Security
Risk Assessment for the Design of I&C Systems in Nuclear Power
Plants”, Nuclear Engineering and Technology, Vol. 44, No. 8, pp.
919–928.
2. Park, J., Suh, Y. (2014), “A Development Framework for Software
Security in Nuclear Safety Systems: Integrating Secure Development
and System Security Activities”, Nuclear Engineering and Technology,
Vol. 46, No. 1, pp. 47–54.
3. Herling, J. (2014), “Advanced Real-Time Manipulation of Video
Streams”, Berlin, Springer, 244 p.
4. Finkle, J. (2013), “U.S. Security Expert Says Surveillance
Cameras can be Hacked”, Reuters, available at: http://www.reuters.com/
article/2013/06/17/us-surveilance-hackers-idUSBRE95G10520130617
5. Gonzalez, R., Woods, R., Chochia, P.A. (2005), “Digital Image
Processing”, translated from Eng., Moscow, Tekhnosfera, 1072 p.
6. CDS-5000 Mini IP Camera Dome System. Nuclear Services,
available at: http://www.rolls-royce.com/~/media/Files/R/Rolls-
Royce/documents/customers/nuclear/po-cds-5000-tcm92-49937.pdf
7. DTR 65 HRC, available at: http://www.ecagroup.com/en/
solutions/dtr-65-hrc
8. Fainzilberg, L.S., Zhuk, T.N. (2009), “Guaranteed Assessment
of Efficiency of Diagnostic Tests Based on Advanced ROC-Analysis”
[Garantirovannaia otsenka effektivnosti diagnosticheskikh testov na
osnove usilennogo ROC-analiza], Control Systems and Machines,
No. 5, pp. 3—13. (Rus)
9. Venttsel, Ye.S. (1999), “The Theory of Probability” [Teoriia
veroiatnostei], Moscow, Vysshaia Shkola, 6th edition, 575 p. (Rus)
10. Bautin, A.V. (2012), “Safety of NPP: from Chernobyl
to Fukushima-1” [Bezopasnost’ atomnykh elektrostantsii: ot
Chernobylia do Fukusimy-1], available at: http://www.agps-2006.
narod.ru/konf/2012/sb-2012/sec-1–12/13–01-12.pdf (Rus)
11. Kobozeva, A.A. (2009), “Basis of a General Approach
to a Problem of Signal Forgery Detection” [Osnovy obshchego
podkhoda k resheniiu problem obnaruzheniia falsifikatsii tsifrovogo
signala], Electrical Machine-Building and Electrical Equipment,
No. 72, pp. 35–41. (Rus)
Received 09.11.2015.
|