Sie sind auf Seite 1von 16

Wireless Pers Commun (2017) 95:475–489

DOI 10.1007/s11277-016-3904-6

Towards Development of a Low Cost Early Fire


Detection System Using Wireless Sensor Network
and Machine Vision

Kehkashan Kanwal1 • Aasia Liaquat2 • Mansoor Mughal2 •

Abdul Rehman Abbasi3 • Muhammad Aamir4

Published online: 24 November 2016


 Springer Science+Business Media New York 2016

Abstract Fire is one of the most prominent threat to safety of human and property, in both
domestic and industrial setups. Efficiently combating a fire threat, usually, depends on how
early the fire is detected. This paper reports work for development of a low cost wireless
sensor-based system for surveillance and early fire detection, using machine vision tech-
nique. The system consists of an on-board camera node, capable of transmitting videos
over wireless network to a remote host computer that runs an image processing based fire
detection algorithm. The system is standalone and portable with the capability of trans-
mitting videos to virtually anywhere in the world. Prototype of the system has been
successfully tested, performing video streaming alongwith segmentation of fire regions
using HSI features of the retrieved images. Future work will inlcude automatic fire
detection and alarm generation alongwith the extension of the system on multiple and
widely scattered transmission nodes.

& Kehkashan Kanwal


Kehkashan.kanwal@hamdard.edu.pk
Aasia Liaquat
aasialiaqat@gmail.com
Mansoor Mughal
mansoormughal07@gmail.com
Abdul Rehman Abbasi
arehman.abbasi@paec.gov.pk
Muhammad Aamir
maamir@ssuet.edu.pk
1
Hamdard Institute of Engineering and Technology, Hamdard University, Karachi, Pakistan
2
NED University of Engineering and Technology, Karachi, Pakistan
3
Design Engineering and Applied Research Lab, KINPOE (Affiliated with PIEAS), Karachi,
Pakistan
4
Sir Syed University of Engineering and Technology, Karachi, Pakistan

123
476 K. Kanwal et al.

Keywords WSN  Early fire detection  Image processing

1 Introduction

Each year, fires result in huge loss of lives and millions of property damage within the
country, and billions worldwide (refer Fig. 1). Since the fire causes serious disaster, a
system that could detect the fire early, is very appealing for both personal safety and
commercial applications [1] .
At present, the conventional fire alarm systems, primarliy based on infrared sensors,
optical sensors and (or) ionization sensors, typically, detect the presence of smoke,
heat, or radiation, using ionization or photometry. The whole system costs a huge
amount of money and it takes a finite time in locating and detecting the presence and
origin of fire. To overcome these shortcomings, video-based fire detection is a suit-
able candidate [2–4].
Healey et al. [5] presented their novel fire detection system exploiting an image-
based approach. In fact, image-based fire detection methods are suitable since they can
provide valuable information such as location, size, and most significantly the extent of
fire.
After acquisition of data, the next significant step is to transmit the data timely and
efficiently. For that matter, wireless sensor network (WSN) being a group of special
sensors, equipped with communication protocols that transmit information sensed by the
sensors over distant locations, are of due consideration. In fact, the WSNs that consist of
multiple detection stations known as sensor nodes, each of which is small, lightweight,
portable and not to mention very economical as well, have been used in a variety of
industries from science and medicine to military and businesses.These sensor networks
measure a wide range of phenomenon including atmospheric temperature, heart rate,
volcanic eruptions, and even the sound of a sniper rifle [6].
This paper presents progress on a prototypic system that would help in provision of
early fire warning at very low cost, by continuously streaming video through WSN from
the surveillance area to a base station, which will function as surveillance centre by
executing image-based algorithm to detect fire from the received videos. Here, we provide
a brief account on state of the art in Sect. 2. In Sect. 3, we discuss hardware implemen-
tation of our approach and results. Finally, in Sect. 4, we conclude and suggest future
work.

Fig. 1 Devastating huge fire spread, leading to heavy loss of lives of about 300 workers at a local garment
factory in Karachi, Pakistan in 2012

123
Towards Development of a Low Cost Early Fire Detection System… 477

2 Literature Review

2.1 Data Acquisition and Transmission

There are two possible approaches for fire detection using wireless sensor network. First is
to use conventional infrared (IR) detectors, ionization detectors and (or) smoke detectors
on sensor nodes and transmit the data to the surveillance centre. Secondly, to use camera as
sensor at the node [7]. Then, image processing can be done either at the node itself and
transmit the detected information to the surveillance centre or transmit the video frames to
the surveillance center for image processing.
Alkhatib [8] used simple and cheap temperature sensors on nodes which were pro-
grammed on a certain threshold temperature, above which the node would send an alarm to
the sink. The target was to realize that whether the fire initiated was normal or wild. This
was expected to lower the possibility of false alarms and was also supposed to provide
some information about behaviour of fire.
Lloret et al. [9] studied the energy consumed by four wireless IP camera devices. The
wireless sensor network was connected to internet through a gateway and the images of the
cameras had been seen from anywhere in the world. The bandwidth consumption mea-
surements given by the test bench showed that the system supported up to 34 wireless IP
cameras in each access point. The design was scalable.
The use of WSN in video signal transmission for various applications has been studied
as well. Kurata et al. [10] did feasibility study of risk monitoring for buildings using the
smart sensor. A smart sensor based on Berkeley Mote platform was introduced. They
investigated performance of mote through shaking table tests. The acceleration sensor was
tested, and its performance for wireless measurement and specific risk monitoring appli-
cations, such as damage detection was assessed. MICA2 Mote was found to be an effective
tool for risk monitoring in buildings.
Al-Marakeby [11] focused on the problems related to WSN image and video trans-
mission and attempted to provide solutions. He replaced the traditional node with camera
node since the traditional sensors such as temperature and pressure sensors, generate
limited amount of data at quiet low rate. Thus, a low-powered device was enabled to
control entire process. For monitoring, detection algorithms were used. After deploying
camera sensor node, the user defines different object location using a graphical user
interface (GUI). For motion detection, three methods were applied, i.e. frame differencing,
optical flow and background subtraction. Hierarchal structures for representing image were
used to reduce memory burden and computational complexity required for motion
detection algorithm.
Stipaničev et al. [12] created a wildfire observer network called Istria iForest Fire Net.
Automatic fire detection and automatic surveillance were done by using geographic
information system (GIS). The system was both, integral and intelligent since it was based
on fusion of different data types and used artificial Intelligence technology from multiagent
organization.
True [13] demonstrated a multi-featured classification approach to detect fire in video
data. Initially, he introduced an algorithm to detect moving regions in the video. In the next
video, fine colored red pixels are extracted using precipitator from the detected moving
regions. The precipitator had two layers, three input nodes; two for each of the color channel
and one node is the hidden layer. In the next step, 50–100 frames are forced and centered each
of these fire moving color regions. In the next step images were sent for dynamic texture

123
478 K. Kanwal et al.

analysis using the statistical property of sequence, auto regressive moving average model
(ARMA) for the tangent image sequence was created. At final stage, a nearest neighbour
algorithm was used to classify dynamic texture model as being fire or not.
Kochláň et al. [14] used an approach for the traffic monitoring based on raspberry-pi
signal bound computer. Video signal processing and analysis was targeted with low-
powered consumption along with reliability. Image information retrieved from raspberry-
pi HD camera module was analysed for moving objects. After classification of objects,
these were sent to noded RF transceiver. Main targetted parameters were volume and
speed. Image processing was performed using GPU. Information stored in SD card was
visualized into encoded live stream and then sent to server.
In the present work, we intend to not only reduce the complexity of above mentioned
schemes but also provide a low cost solution for easy implementation of an early fire
detection system.

2.2 Image Processing of Fire Data

In a fire incidence, if the initial combustion could be detected in time, not only the fire can
be extinguished immediately but the damage may be reduced greatly.
Today, most of the fire or flame detection techniques are usually based on smoke
sampling [15, 16], temperature sampling [15–18] and air transparency testing, in addition
to the traditional ultraviolet and infrared flame detectors [15–19]. However, most of these
detectors suffer from some severe problems. First, they require a close proximity to the fire
source. In addition, they are not always reliable because they do not always detect the
combustion itself. Instead, they detect the by-products of combustion, which may be
produced in other ways. Therefore, they usually result in higher false rates.
In order to improve the fire detection rate, experts [15, 16, 20, 21] mixed the earlier
mentioned techniques by adopting hybrid sensors. Others [16, 22–24] have also applied soft
computing techniques, to assist such detection. However, all of these methods seldom
provide additional descriptive information about location, size, and extent of fire. In addition,
once fire has been detected by the traditional detectors, it usually has become very large. That
is, these detectors usually fail during initial combustion and thus may result in greater
damages.
Healey et al. [5] reported their novel fire detection system, based on image approach.
Since then, many image-based fire detection methods have been proposed because images
can provide more information regarding location, size, and extent of fire. In fact, Healey
et al. [5] presented an aircraft jet fuel fire detection system using a color video as input on a
pre-divided scene to approximately locate fire and estimate the flame size by detecting the
grids that appear in video frames.
Noda and Ueda [25] used infrared images to detect tunnel fires by comparing the normal
intensity histograms of various tunnel fire images with that of a captured image to
determine whether there exists an obvious intensity variation.
Foo [26, 27] applied statistical measures, such as first moment, standard deviation, and
mean, to estimate flame sizes of aircraft dry bays and engine compartments using gray-
level images by analyzing the relationships of the intensity changes between different fire
burning degrees from a set of test flame images. The above two approaches were designed
for specific environments and could not detect spurious flames because they use gray-level
images as inputs.
Yamagishi and Yamaguchi [28, 29] utilized a neural network to determine fires from
Fourier transforms of fire contours extracted from color images based on the HSI color

123
Towards Development of a Low Cost Early Fire Detection System… 479

space. Later, Phillips et al. [30] proposed a sophisticated method for recognizing flames in
color video by considering the temporal variation of flames.
The above mentioned methods are extensive in distinguishing between real fires and
spurious fires. However, they are too complex for real-time applications. In addition, they
do not provide the detection of initial combustion.
In order to have an image-based flame detection monitoring system to be of practical
use so that users would be informed with a fire alarm as early as possible and they could
view the flame images on displays, the system must work in real-time and must be applied
to more general environments instead of being restricted to some specific situations. The
current work has been done using video signal transmission from sensor node to the
surveillance centre and processing at the base station to address the mentioned problems.

3 Hardware Implementation and Results

3.1 Data Acquisition and Transmission

Block diagram of our proposed system is shown in Fig. 2. The hardware components used
in the project are listed below and shown in Fig. 3.

Fig. 2 Block diagram of the proposed system

123
480 K. Kanwal et al.

• Raspberry pi Camera (Sensor)


• Raspberry Pi (Controller)
• Battery (Power Source)
• Internet device (Communication Protocol)
Selection of ZigBee was ruled out at the very first stage because the data rate required
for video transmission cannot be handled by it and alternative methods were explored.
The camera is interfaced with controller (raspberry pi 2—model B) and used to con-
tinuously record the video. The system is powered up by a portable power bank of 10,000
mAh. The internet device is a portable 4G wifi device. The system uses a USB adapter and
is capable of connecting to any given wifi network. Also the controller can use any
WebCam or Surveillance camera so that total cost of the system is manageable as per our
intended goal.
The Raspberry pi is used to transmit the video. There are many video streaming pro-
tocols freely available and any of them can be used. The aim is to stream accessible
anywhere so that surveillance can be done remotely. The use of IP camera was avoided to
further reduce the cost down and also to make it possible to integrate the suggested system
with already deployed security surveillance system.
Figure 4 presents screen shot of the raspberry pi when ustream code is executed
whereas Fig. 5 shows screenshots of different video streams that were captured. These
screenshots were taken on three different internet browsers at various locations. The stream
has approximate time delay of 20–30 s which is acceptable for the given application.

3.2 Image Processing Sub-system

Based on the deficiencies of existing early fire detection limitations, a vision-based fire
detection system is proposed. The proposed algorithm processes visual information

Fig. 3 Hardware components used in the proposed system (top left and right) are Raspberri Pi board and
Camera. (Bottom left and right) are Wifi adapter and source device

123
Towards Development of a Low Cost Early Fire Detection System… 481

Fig. 4 Code for transmitting video at transmission end

acquired through WSN as mentioned in previous section. This work is aimed to analyse
image sequences, i.e videos.
Since the classifier in this work is supposed to differentiate between fire and non-fire
only, the actual task was to separate fire objects from non-fire. The first task was to
generate a database of fire videos and non-fire videos. Since it was not practically possible
to have an experimental setup to have the camera placed and have fire events, videos from
internet were used to generate sample fire and non-fire database. From the videos,
screenshots were taken and pre-processed in Adobe Photoshop. Then they were analyzed
in MATLAB for their corresponding HSI and RGB values. These HSV values were then
used in designing an Image Processing algorithm using C?? in Microsoft Visual Studio.

3.3 Database Generation

From all the downloaded videos, 30 clips of both fire and non-fire scenes were generated,
each was 30 s long. The trimming of fire videos and preparation for HD display was done
in Windows Movie Maker (Version 2012). Each of this clips were processed in video to
image (jpg) converter software. Figure 6 shows some frames that were obtained.
From all of the obtained screenshots 100 fire patches and 50 non-fire patches were
obtained using Adobe Photoshop. Each image was loaded and the fire part was cropped
using Lasso tool and copied to new blank window. The back ground of the image was
removed to have the selected image only. Some examples of obtained patches are shown in
Fig. 7.

123
482 K. Kanwal et al.

Fig. 5 Snapshots of video streaming from web browser at different location, (top left and right) show video
streaming from smart phone and laptop respectively on Google Chrome, (bottom) shows video streaming on
laptop using Mozilla Firefox

3.4 Feature Extraction

All of these fire patches were analyzed in MATLAB 2013. Each image was converted into
HSI image first and then mean of their H, S and I values was calculated. All the HSI values
were stored in a single matrix. The mean of hue, saturation and intensity is simple linear
mean because MATLAB presents normalized H,S and V. All the HSI values were again
averaged to get mean HSI values for fire. These values of fire are used in image processing
algorithm for image segmentation.
For OpenCV 3.10, the range of saturation and intensity are 0–255 hence the normalized
value was multiplied with 255. For Hue it is 0–179. The normalized value from matlab
were first converted into angular value (0–360) and then to 0–179.

3.5 Image Processing Algorithm

For image processing algorithm, openCV 3.10 library was used with IDE of Microsft
Visual studio and coding language employed is C??. The code works as follows
• Reads a video file from directory in which code is saved.
• Convert the video frames in HSI color space.

123
Towards Development of a Low Cost Early Fire Detection System… 483

Fig. 6 Grid view of obtained frames from video to jpg converter

• Do image segmentation.
• For this, the algorithm looks for each image pixel that lies within 10 intensity value
from mean HSI value of fire obtained from feature extraction phase.
• All pixels which lie between the range are assigned value of 1 (bright pixel or
white) and all others are assigned value of 0 (dark pixel or black).
• After the binary image segmentation, the segmented image is used to perform
morphological operation of opening (erosion followed by dilation), this step makes the
boundaries of fire like image visible.
• After morphological operations have been done, the thresholded video is displayed.
• Following morphological operations, the algorithm finds contours in the segmented
video feed. It retrieves all of the contours and reconstructs a full hierarchy of nested
contours. It uses method that compresses horizontal, vertical, and diagonal segments
and leaves only their end points.
• The algorithm finds hull for each contour it has retrieved.

123
484 K. Kanwal et al.

Fig. 7 Sample fire patches

• Whenever the algorithm successfully finds any contour, it draws contour around the
pixel location where it has find fire and prints fire detected on original camerafeed
screen.

3.6 Discussion on Results

Figures 8, 9 and 10 show results when above mentioned algorithm was implemented.
Every time, when the code finds fire pixels in video feed, it outputs fire detected on
the original video. However, like most of the practical systems, it has limitations
regarding creation of false alarm when a fire like color object is detected in the image
frame.
The strength of the system lies in the speed of the system. It is faster than smoke particle
detectors, especially, if the fire is at a distant location. Secondly, the range of the system is
high. The algorithm works on the entire screen of the video, the range of fire detection
depends on the range of which the camera can capture. It is especially useful in large open
areas, such as an open field. The key outcome is that the system can be used to stream
video to virtually any part of the world without using IP camera with acceptable time
delay.

123
Towards Development of a Low Cost Early Fire Detection System… 485

Fig. 8 Manual segmentation using HSI features and contour formation of fire regions from training samples

Fig. 9 Algorithm outputs fire detected and draws contour around the detected patch

123
486 K. Kanwal et al.

Fig. 10 Result of input image sequence when there is fire patch in frame. Warning message showing
detection of fire

4 Conclusion and Future Work

The report presented here is part of an ongoing developmental work for low cost devel-
opment of an early fire detection and surveillance system using machine vision based
techniques. The outcome of the proposed scheme is promising and more specifically,
attractive for economical implementation in domestic context. The system has some
limitation, i.e. it gives false alarm in some cases when there exists a highly similar fire-like
colored object. Especially, since the input videos have variable illumination that might
have made feature extraction erroneous.
A key extension of the present work is to integrate the fire detection system with already
installed CCTV monitoring systems that are nowadays in ample use.

References
1. Çetin, E. A., Dimitropoulos, K., Gouverneur, B., Grammalidis, N., Günay, O., Habiboǧlu, Y. H., et al.
(2013). Video fire detection—review. Digital Signal Processing, 23, 1827–1843.
2. Wang, S., & Chen, J. (2012). A method of video flame detection based on multi-feature fusion. Journal
of Convergence Information Technology, 7(21), 634–642.
3. Kim, Y.-H., Kim, A., & Jeong, H.-Y. (2014). Rgb color model based fire detection algorithm in video
sequences on wireless sensor network. International Journal of Distributed Sensor Networks, 10,
1477–1550.
4. Shen, P., Zhang, L., Song, J., Xu, H., Zhou, L., Wei, W., et al. (2013). Dm642-based fire detection in
video sequences using statistical color model. International Journal of Digital Content Technology and
its Applications, 7(3), 669–678.
5. Healey, G., Slater, D., Lin, T., Drda, B., & Goedeke, A. D. (1993). A system for real-time fire detection.
In Proceedings of the IEEE conference on computer vision and pattern recognition.
6. Hempstead, M., Lyons, M. J., Brooks, D., & Wei, G.-Y. (2008). Survey of hardware systems for
wireless sensor networks. Journal of Low Power Electronics, 4(1), 1–10.
7. Aslan, Y. E., Korpeoglu, I., & Ulusoy, Ö. (2012). A framework for use of wireless sensor networks in
forest fire detection and monitoring. Elsevier Journal on Computers, Environment and Urban Systems,
36, 614–625.

123
Towards Development of a Low Cost Early Fire Detection System… 487

8. Alkhatib, A. A. A. (2013). Smart and low cost technique for forest fire detection using wireless sensor
network. International Journal of Computer Applications, 81(11), 12–18.
9. Lloret, J., Garcia, M., Bri, D., & Sendra, S. (2009). A wireless sensor network deployment for rural and
forest fire detection and verification. Sensors, 9, 8722–8747.
10. Kurata, N., Spencer Jr., B. F., & Ruiz-Sandoval, M. (2004). Building risk monitoring using wireless
sensor network. In Proceedings of the 13th world conferences on earthquake engineering.
11. Al-Marakeby, A. (2013). Camera-based wireless sensor networks for e-health. International Journal of
Advanced Research in Computer and Communication Engineering, 2, 4757–4761.
12. Stipaničev, D., Štula, M., Krstinić, D., Šerić, L., Jakovčević, T., & Bugarić, M. (2010). Advanced
automatic wildfire surveillance and monitoring network. In Proceedings of the 6th international con-
ference on forest fire research.
13. True, N. (2016). Computer vision based fire detection. Online available at http://www.cseweb.ucsd.edu/
classes/wi09/cse190-a/reports/ntrue.pdf. Accessed March, 2016.
14. Kochláň, M., Hodoň, M., Čechovič, L., Kapitulı́k, J., & Jurečka, M. (2014). WSN for traffic monitoring
using raspberry piboard. In: Proceedings of the federated conference on computer science and infor-
mation systems.
15. Lin, S. Y., Luo, R. C., & Su, K. L. (2003). A multi agent multi sensorbased security system for
intelligent building. In Proceedings of the IEEE conference on multi sensor fusion and integration for
intelligent systems.
16. Muller, H. C., & Fischer, A. (1995). A robust fire detection algorithm for temperature and optical smoke
density using fuzzy logic. In Proceedings of the 29th international conference on security technology
(pp. 197–204).
17. Bartoloni, A., Cisbani, E., Marchese, M., Efisei, G., & Salvati, A. (2002). Early fire detection system
based on multi-temporal images of geostationary and polar satellites. In Proceedings of the IEEE
international geoscience and remote sensing symposium.
18. Odic, R. M., Jones, R. I., & Tatam, R. P. (2002). Distributed temperature sensor for aeronautic
applications. In Optical fiber sensors conference technical digest.
19. Arrue, B. C., Ollero, A., & Matinez de Dios, J. R. (2002). An intelligent system for false alarm reduction
in infrared forest-fire detection. IEEE Intelligent Systems and Their Applications, 15(3), 64–73.
20. Kaiser, T. (2000). Fire detection with temperature sensor arrays. In Proceedings of the 34th interna-
tional conference on security technology.
21. Luo, R. C., Su, K. L., & Tsai, K. H. (2002). Fire detection andisolation for intelligent building system
using adaptive sensoryfusion method. In Proceedings of the international conference on robotics and
automation.
22. Neubauer, A. (1997). Genetic algorithms in automatic fire detection technology. In Proceedings of the
2nd international conference on genetic algorithms in engineering systems: Innovations and
applications.
23. Xiao, J.-M., & Wang, X.-H. (2003). A fuzzy neural network approach to fire detection in ships. In
Proceedings of the 12th IEEE international conference on fuzzy systems.
24. Sun, C. T., Jang, J. S. R., & Mizutani, E. (1997). Neuro fuzzy and soft computing. Upper Saddle River:
Prentice-Hall.
25. Noda, S., & Ueda, K. (1994). Fire detection in tunnels using animage processing method. In Pro-
ceedings of the IEEE conference on vehicle navigation and information systems.
26. Foo, S. Y. (1995). A machine vision approach to detect and categorizehydrocarbon fires in aircraft dry
bays and engine compartments. In Proceedings of the IEEE conference on industry applications.
27. Foo, S. Y. (2000). A fuzzy logic approach to fire detection in aircraft dry bays and engine compart-
ments. IEEE Transactions on Industrial Electronics, 47, 1161–1171.
28. Yamagishi, H., & Yamaguchi, J. (1999). Fire flame detectionalgorithm using a color camera. In IEEE
international symposium on micromechatronics and human science.
29. Yamagishi, H., & Yamaguchi, J. (2000). A contour fluctuation data processing method for fire flame
detection using a color camera. In Proceedings of the 26th IEEE annual conference on industrial
electronics society.
30. Phillips, W., Shah, M., & Da Vitoria Lobo, N. (2000). Flame recognition in video. In IEEE workshop on
applications of computer vision.

123
488 K. Kanwal et al.

Kehkashan Kanwal has done her B.E. in Biomedical Engineering and


M.Engg. in Mechatronics Engineering from NEDUET, Karachi. Cur-
rently she is serving as Lecturer in Hamdard Institute of Engineering
and Technology. Her major interest areas are: Digital Image Pro-
cessing, Pattern Recognition, Machine Vision, Rehabilitation Engi-
neering and Biomedical Instrumentation.

Aasia Liaquat has done her M.Engg. in Mechatronics Engineering


from NEDUET, Karachi. Her major intereset area is Digital Image
Processing.

Mansoor Mughal is doing his M.Engg. in Mechatronics Engineering


from NEDUET. His research interests are: Assistive Devices and
Rehabilitation Engineering, Brain Computer Interface (BCI), Near
Infra-Red Spectroscopy (NIRS), EEG and Digital Image Processing.

123
Towards Development of a Low Cost Early Fire Detection System… 489

Dr. Abdul Rehman Abbasi holds a Ph.D. in Mechatronics Engi-


neering from the Asian Institute of Technology, Thailand. Earlier, he
completed his Masters and Bachelors in Electrical Engineering from
NED University of Engineering and Technology, Karachi. Before
proceeding Thailand for Ph.D. studies in 2005, he joined Pakistan
Atomic Energy Commission and worked at Karachi Nuclear Power
Plant for 6 years as shift operation engineer. Currently, he is working
as Head of Master of Engineering Program and In-charge, Design
Engineering and Applied Research Laboratory, Karachi Institute of
Power Engineering, Karachi, which is an affiliated Institute of Pakistan
Institute of Engineering and Applied Sciences. He has more than 27
research publications in various journals and international peer-re-
viewed conferences of high repute. His research interests are to
Explore the Frontiers in Sensing and Actuation, Digital Image Pro-
cessing, Machine Vision and Applied Robotics for Industry Problem
Solutions.

Dr. Muhammad Aamir was born in Karachi Pakistan, on July 3,


1976. He received M.S. degree in Electronic Engineering (with spe-
cialization in Telecommunication) in 2002 and B.S. in Electronics
Engineering in 1998. He accomplished his Ph.D. in Electronic Engi-
neering from Mehran University of Engineering and Technology.
During his Ph.D. studies, he accomplished his research work at the
University of Malaga under Erasmus Mundus Scholarship. He has
authored and co-authored more than 30 research papers published in
various journals and conferences of international repute. He is a life
member of Pakistan Engineering Council and professional member of
IEEE for last 12 years. He was also awarded with a grant by the
Ministry of Education Spain to teach at the University of Malaga
which he availed in May 2012. He is also Member of NCRC in the
Higher Education Commission (HEC) for revision of Electronic
Engineering Curriculum and Telecommunication Engineering cur-
riculum at the national level. He is guest editor for special issue of
Springer’s Journal with title ‘‘Wireless Personal Communication’’ which had published in May 2015. He is
also HEC approved supervisor for Ph.D. candidates.

123
Wireless Personal Communications is a copyright of Springer, 2017. All Rights Reserved.

Das könnte Ihnen auch gefallen