Video Surveillance Device Failure Detection by IoT

Full text

Video Surveillance Device Failure Detection by IoT
Mohana Puneeth
School of C&IT
REVA University
Bangalore, India
pune269@gmail.com
Geetha D. Devanagavi
School of C&IT
REVA University
Bangalore, India
dgeetha@reva.edu.in
Nikhath Tabassum
School of ECE
REVA University
Bangalore, India
nikhath.tabassum@reva.edu.in
Abstract— The main objective of Video Surveillance is to
traverse back for any untoward incident and to investigate for
the actual cause or reason. In the current scenario, CCTVs
(Closed Circuit Televisions) are placed all over for recording the
events but it is very difficult to monitor the working of these
CCTV (Closed Circuit Televisions) cameras in real time. It
requires constant manual monitoring to realize if a camera is not
working by looking at the recording screen. This research study
proposes a novel method to find faulty cameras by analyzing
whether a camera is connected to the internet network or not by
using the ICMP (Internet Control Message Protocol) echo
command method to a given server and its reply packets can be
used to identify the faulty devices connected to a given network.
The turnaround time for manual checking of faulty cameras and
raising concern can be reduced drastically by saving more time
and most importantly the video surveillance keeps tracking
events all the time.
Keywords— NVR/DVR (Network or Digital Video Recording),
Video Surveillance, Cyber Security, CCTV (Closed Circuit
Televisions), ICMP (Internet Control Message Protocol), Firewall.
I. INTRODUCTION
The perseverance of surveillance camera is to detect, track
and classify the target of the surveillance area or the region.
The important questions remain’s to be addressed on any
video surveillance device as an effective tool for crime
prevention, crime resolution and crime protection is if its
working or not!!
The system admin is caught off guard if someone asks if the
video surveillance devices connected to a given network is
working or not in real time and to add on the admin cannot use
the camera to change the angel, zoom for more details or video
recording might be stopped all-together cause the device is
offline. This are the few actions the camera fails to do. With
available of highspeed internet and various video device
manufacturer it is very easy to find the NVR/DVR ( Network
or Digital Video Recording) into market. NVR/DVR are prone
to attacks based on the software which is open to outside world
and lacks safety features.
Currently in market cyber security threat prevention software
are rolled out by each manufacturer, in mega company camera
devices are purchased from different manufacturers/vendors
cause of video surveillance devices are used internal(inside the
building) or external(outside building) and this cannot be
combined and tracked in a single point of server.
The solution is thus by placing video devices channel in a
given organization is by deleting the IP details received from
the manufacturers and re-writing the new IP address for each
video surveillance device and placing them into a company
given server along with gateway and subnet assigned to it, so
that it will create a firewall for any external devices to get
inside the server and with help of a software by consistently
pinging ICMP (Internet Control Message Protocol) echo IP’s
along with the network and store their responses for those
devices which are not reverting the signal, thus this will lead
to the device IP’s failure. These failed devices are classified as
offline devices and post that network team will lead the action
item to solve this immediately.
The positive outcome are :
1) Video Surveillance device will be always active and
recording the incidents in real time and this can be
viewed with help of internet at any place and time.
2) With evolving in new technology and threats in cyber
security space, by placing the firewall which is
managed by the company’s internal networking team
will prevent more insights to external viewers and
protect the data without been exposed to outside
world.
3) Most of the manual effort will be reduced by tracking
the device status in real time and thus cutting the time
and more important the data breach and personal
breach restricted zone.
The remainder section is organized as follows. Section II has
literature survey, section III illustrates how the surveillance
camera are connected, section IV has results and discussions,
section V has conclusion.
II. LITERATURE SURVEY
Multiple sensors to identify an object from multiple angles,
obtain additional dimensional information about the object,
and combine multidimensional information[1], anomaly
detection in videos for video surveillance application gives
assured results in regards to real-time scenarios[2], intelligent
video surveillance system is added in the original operation
and maintenance monitoring platform[3], encoding the
surveillance video frames as per the generated salient-map, the
default available HEVC test model is used[4], various
anomality in capturing videos and its solution is captured here
[5], vulnerabilities and types of attach on cyber security in
Proceedings of the Second International Conference on Applied Artificial Intelligence and Computing (ICAAIC 2023)
IEEE Xplore Part Number: CFP23BC3-ART; ISBN: 978-1-6654-5630-2
978-1-6654-5630-2/23/$31.00 ©2023 IEEE 1448
2023 2nd International Conference on Applied Artificial Intelligence and Computing (ICAAIC) | 978-1-6654-5630-2/23/$31.00 ©2023 IEEE | DOI: 10.1109/ICAAIC56838.2023.10140539
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on June 15,2023 at 03:13:12 UTC from IEEE Xplore. Restrictions apply.

video surveillance[6], threats and attacks in IoT for the video
surveillance and its impact on the users [7], machine learning
and classification of port scanning by using data set[8], Visual
image detection in maritime weather condition[9], image
processing by matching algorithm to find the best result[10],
automatic video analyser to find the object[11], two-stream
convolutional networks model for anomaly detection in
surveillance videos[12], multi-object detection utilizing the
Cyber secure Probabilistic Gaussian Mixture Model and
background suppression and another stage is multiple moving
objects tracking utilizing Kernel convoluted moving window
with Kalman filter[13], cybersecurity and cyber risk
management with a focus on data availability[14],
[15] estimating the spatial distribution of surveillance
cameras
There are noticeable different manufacturers/vendors devices
are connected to various network/servers and to check its
status becomes even more cumbersome at one point of
contact.
The importance of solving this issue makes huge importance
in running video surveillance device camera 24/7 and giving
security cover for any un-towards incident which might have
happened or about to happen.
III. ILLUSTRATES HOW THE SURVEILLANCE
CAMERAS ARE CONNECTED
An organization purchases camera from different vendors
for the specific needs. For instance, the difference can be
differentiated with indoor or outdoor, coaxial, or wireless. As
a result, the organization ends up procuring various vendors
for its usage.
So, the first thing an organization should do is to change the
IP address which suits the organization security protocol and
align it to a given a network so that the vendor cannot track the
devices from its base server.
Any new patches or the certificate will be tested first by
retrieving the certificates through "Certification Authority
Proxy Function (CAPF) Information" from the camera
configuration page. The CAPF operation team should
physically check the device working condition before it is
been assigned for installation post IP address and other setting
been changed.
Fig 1: Surveillance connected to a given server.
Fig 1, describes various camera outlet connected to a given
server for example Security Camera, Surveillance Camera wi-
fi camera are in turn connected to a server and that server is
connected to an Internet.
a. Methodology to conceptulize the faulty camera
detection
Fig 2: Faulty Camera Detection Algorithm.
Fig 2: Faulty camera Automation in Building Technology
algorithm workflow is been described here. The camera which
is connected to server via IP based are fetched into a given
environment, post that PING command will interact with
cameras consistently so that a proper communication is
Proceedings of the Second International Conference on Applied Artificial Intelligence and Computing (ICAAIC 2023)
IEEE Xplore Part Number: CFP23BC3-ART; ISBN: 978-1-6654-5630-2
978-1-6654-5630-2/23/$31.00 ©2023 IEEE 1449
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on June 15,2023 at 03:13:12 UTC from IEEE Xplore. Restrictions apply.

maintained all the time. The devices which are offline are
captured into a DB log and it is used for further analysis.
Business Intelligence tool is connected for the future analysis.
The design of the working model can be established with
the below 3 models:
1) To establish if a camera is connected to a given
network or not, this becomes threat to the organization
by not only being offline and missing important
relevant data needed at times but prone to cyber-attack
in near future and it will impact entire organization to
pay the hefty price and more importantly the brand
image will be lost. This can be established by setting
an IP address to the devices manually by entering the
string into algorithm and converting the string into
valid IP address. This will prevent the
vendor/manufacturer to access our application data and
gain useful insights at the backend and establishing the
connection to an organization server only.
2) By constantly invoking the ICMP (Internet Control
Message Protocol) echo command. With the help of
python programming, we can invoke all the IPs of
surveillance devices connected to a given server at
any given point of time thus the target host receives
the echo request and target hosts will respond by
sending an echo reply packet. With this we will be
able to generate the offline devices not connected to
the servers and capturing those data as a log will give
more insight.
3) The loss of signal (the target host receives the
echo request) as a log and this is captured into SQL
(Structure Query Language) database for future
analysis. The ICMP echo response packets will be
saved in a log and these are called offline devices.
b. Tabular below describes the difference between
the old and new method for setting up the video
devices and its new features.
S.no
Old Method
New Method
1
Each device
manufacturers are
connected to
specific server; thus,
the server count and
maintenance cost is
high.
Only one server can be
connected to N device
manufacturers, thus
reducing the servers
deployed, maintenance
and running cost.
2
IPs are set from the
device
manufacturers
IPs are rewritten and
hidden form
manufacturers
3
Device data are
tracked by the
manufacturers
Manufactures cannot
gather the data from
device.
4
Certification/patches
are installed in each
server one at a time.
Certification/patches
needs to be installed at a
single shot going forward.
c. Constraints faced while expirementing.
Could not find any single method on automating all the
different patches or new certification at one go for different
manufacturers, the security clearance and patch clearance
from the vendor needs to be tested and approved individually
and so it leads to uploading the patches individually.
IV. RESULTS AND DISCUSSION
The result will be simulated with the help of dataset prepared
and fed into the system for the checking for online and offline
devices:
i. Data Set:
Table 1 describes the data used for analyzing the device
failure in an environment. Data is been prepared for company
named ABC which is connected to a server, the key
parameters are the IP( which is unique address), Gateway (
which may be unique or not), Address column contains the
device installed place, City column gives the information
about the city, State column gives the information about the
state/provinces and Postal Code column gives about the pin
code. These fields are used are for reference where and all
devices has been installed and if any untoward incident
happens a vendor can physically go this site and replace or fix
the camera. Below sample forecast the data and the
representing fields.
Table 1: Sample data set used for simulating the result
ii. Python Code:
Data from the server end is read by a Python code, initially the
IP address will be in a string, that converts to a valid unique IP
address for the system to consume later the ICMP (Internet
Control Message Protocol) echo command is checked for all
IPs inside the server and echo response packets will be saved
in a log and these are called offline devices.
A. Business Intelligent tool (Tableau) is used to find the
Best result of given data set.
Once the data is exported after attaining the device status
as online or offline, Business Intelligence tool will build the
dashboard, this will analyse the data and to give more insights
to the network teams and to end users in a single window
without much difficulty.
Proceedings of the Second International Conference on Applied Artificial Intelligence and Computing (ICAAIC 2023)
IEEE Xplore Part Number: CFP23BC3-ART; ISBN: 978-1-6654-5630-2
978-1-6654-5630-2/23/$31.00 ©2023 IEEE 1450
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on June 15,2023 at 03:13:12 UTC from IEEE Xplore. Restrictions apply.

The same loss of signal data is been transformed to networking
team to analyse the faulty devices and below action will be
taken by team to troubleshoot.
i. How to view the best result.
Dashboard gives the total number of devices which are
connected to a given network, followed by online device count
and offline device’s overall, this calculate the percentage of
number of devices which are offline and action can be taken
accordingly. The dashboard can be drilled down to each state
level to get more insight.
To achieve the best result, always the camera needs to be
active and running at the back end so that number of offline
will be reduced to 0.
Table 2: Screenshot of tableau dashboard taken.
Table 2 dashboard shows the ABC company device status in
real time with various parameters (Address, state, zip, Ip,
status and installation date) to gauge the offline devices life.
B. Network Troubleshooting steps
Any surveillance camera should work within range of 12-24V
DC, if slight change in voltage fluctuation will result in
devices being offline, so voltage parameter to be in key factor.
on the IP default, at times devices might get the IP to be set as
default. If a device certificate is expired, a new certificate
needs to be installed in the server for the device to be up and
running.
If above all process does not able to troubleshoot the devices
which are offline, that needs to be replaced with help of a
technician physically visiting the location and send the
feedback for any.
Fig 3: Networking team troubleshooting.
Fig 3. describes network team is handy with surveillances
devices data which are offline, they must start working on
troubleshooting by checking the LAN cable connectivity with
various GATEWAY channels are enabled or not.
V. CONCLUSION
Video surveillance is a part of human life in our daily activity,
to keep an eye on its working condition and this becomes a
tedious task when it comes to a large organization where day
in and day out activity needs to be recorded for quality
purpose. To automate the process of monitoring the offline
vvideo surveillance devices with latest technology is
important and reduces lot of human effort and cost in running
the process. By implementing the BI (Tableau) will give more
pictorial representation of all devices and its status on any
given point of time. This proposed method prevents from
cyber-attack and phishing into an organization where lot of
surveillance is active secure zone. This proposed method will
increase security as well as instant monitoring of the devices
when its offline.
REFERENCES
[1] Xiang Li and Yu Fu, “Design and Research of Computer Network
Education and Teaching System Based on Multisensor Information Fusion”,
Hindawi Mobile Information Systems Volume 2022, pp.3, 2022.
[2] Mohana Mohana, Ruben J Franklin, Vidyashree Dabbagol, “Anomaly
Detection in Videos for Video Surveillance Applications using Neural
Networks”, DOI: 10.1109/ICISC47916.2020.9171212, pp. 632-638,2020
[3] Chuyun She, Runlan Zheng, Qiaoyin Yang and Shaowei Liang, Research
of Intelligent Video Surveillance System based on Artificial Neural Network,
Conf. Ser. 2181 012057, pp. 3, huyun She et al 2022 J. Phys.: Conf. Ser. 2181
01205
[4] Asma Zahra, Mubeen Ghafoor, Kamran Munir,· Ata Ullah and Zain Ul
Abideen, “ Application of region based video surveillance in smart cities using
deep learning”, https://link.springer.com/article, 2021, p.18-22,2021.
[5] K. Nithesh, N. Tabassum, D. D. Geetha and R. D. A. Kumari, "Anomaly
Detection in Surveillance Videos Using Deep Learning," 2022 International
Conference on Knowledge Engineering and Communication Systems
(ICKES), Chickballapur, India, 2022, pp. 1-6, doi:
10.1109/ICKECS56523.2022.10059844.
[6] Preethi Vennam, Pramod T. C, Thippeswamy B. M., Yong-Guk Kim and
Pavan Kumar B. N. “Attacks and Preventive Measures on Video Surveillance
Systems: A Review“ , applied science, vol 11, issue 12, pp. 3.10, 2021
[7] Naor Kalbo , Yisroel Mirsky , Asaf Shabtai and Yuval Elovici , “The
Security of IP-based Video Surveillance Systems“, research gate, pp. 3-8,
2020
[8] Muhammad Aamir1a, Syed Sajjad Hussain Rizvi1b, Manzoor Ahmed
Hashmani2 , Muhammad Zubair3 , Jawwad Ahmad4, “Machine Learning
Classification of Port Scanning and DDoS Attacks: A Comparative Analysis
“, Mehran University Research Journal of Engineering and Technology, Vol.
40, No. 1, 215- 229, January 2021, pp. 215-221, 2021
[9] Moosbauer, Sebastian, Daniel Konig, Jens Jakel, and Michael Teutsch. “A
Benchmark for Deep Learning Based Object Detection in Maritime
Environments”, In Proceedings of the IEEE Conference on Computer Vision
and Pattern Recognition Workshops, 2019.
[10] Nehashree M R et.al., “Simulation and Performance Analysis of Feature
Extraction and Matching Algorithms for Image Processing Applications”
International Conference on Intelligent Sustainable Systems (ICISS-2019).
[11] Kong, Lingchao, Ademola Ikusan, Rui Dai, and Jingyi Zhu. “Blind Image
Quality Prediction for Object Detection”, In 2019 IEEE Conference on
Multimedia Information Processing and Retrieval (MIPR), pp. 216-221.
IEEE, 2019.
Proceedings of the Second International Conference on Applied Artificial Intelligence and Computing (ICAAIC 2023)
IEEE Xplore Part Number: CFP23BC3-ART; ISBN: 978-1-6654-5630-2
978-1-6654-5630-2/23/$31.00 ©2023 IEEE 1451
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on June 15,2023 at 03:13:12 UTC from IEEE Xplore. Restrictions apply.

[12] Wangli Hao,1Ruixian Zhang,1Shancang Li,2Junyu Li,1Fuzhong
Li,1Shanshan Zhao,2and Wuping Zhang1, “Anomaly Event Detection in
Security Surveillance Using Two-Stream Based Model”, Hindawi , Volume
2020 | Article ID 8876056 | https://doi.org/10.1155/2020/8876056, pp. 3-6,
2020.
[13]Kalyanapu Srinivas a, Laxman Singh b, SubbaReddy Chavva c, Bhasker
Dappuri d, Saravanan Chandrasekaran e, Shamimul Qamar f, “ Multi-modal
cyber security based object detection by classification using deep learning and
background suppression techniques” , Science Direct, Volume 103, pp.
October 2022, 108333
[14] Frank Cremer · Barry Sheehan · Michael Fortmann· Arash N. Kia·Martin
Mullins · Finbarr Murphy · Stefan Materne , “ Cyber risk and cybersecurity:
a systematic review of data availability”, springernature, 41288, 17 February
2022, pp. 2-4, 2022.
[15] Hao Sheng, Keniel Yao ,Sharad Goel, “Surveilling Surveillance:
Estimating the Prevalence of Surveillance Cameras with Street View Data”
AIES '21: Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and
Society, pp. 221, July 2021
Proceedings of the Second International Conference on Applied Artificial Intelligence and Computing (ICAAIC 2023)
IEEE Xplore Part Number: CFP23BC3-ART; ISBN: 978-1-6654-5630-2
978-1-6654-5630-2/23/$31.00 ©2023 IEEE 1452
Authorized licensed use limited to: REVA UNIVERSITY. Downloaded on June 15,2023 at 03:13:12 UTC from IEEE Xplore. Restrictions apply.