International Journal of Science and Research (IJSR)

ISSN (Online): 2319-7064

Index Copernicus Value (2013): 6.14 | Impact Factor (2015): 6.391

Intelligent Traffic Control System using Image

Processing
Parichita Basak1, Ramandeep Kaur2
1, 2
Guru Gobind Singh Indraprastha University, Bharati Vidyapeeth’s College of Engineering, Department of Information Technology,
A-4, Paschim Vihar, Rohtak Road, New Delhi-110063

Abstract: Traffic congestion is becoming more and more serious day by day. Main reasons for augmented traffic jam are increasing
number of vehicles, the poor infrastructure and no proper distribution. The main reason for traffic is increased number of vehicles and
increased number of population and development of country as whole. Management of traffic in India is a tough job and only manual
efforts cannot solve this serious issue. We need a system to handle this situation more effectively. We need a dynamic system that is
capable of controlling traffic as well as avoid congestion of roads, known as Intelligent Traffic Control System. In this project, we are
dealing with traffic via image processing. Using different types of image processing algorithms, the vehicles can be detected on basis of
which traffic lights can switch.

Keywords: Intelligent traffic control, traffic lights, image processing, traffic control, traffic congestion

1. Introduction sends the controller sensitive to signals. The system will

work only if the assumed flow matches the actual traffic flow.
Detection and classification of vehicle is essential for
effective traffic control, for which, traffic related information Manual Controlling [5] requires extra man power. Moreover,
needs to be collected and analyzed.. Some other techniques only skilled operators can make judgements, hence, there is
are Magnetic Wireless Sensor detectors[1], Radio very high work load on skilled operators.
Frequency[2], Regression Analysis[3], Motion Vector
Technique[4], etc. Standard traffic control system is Manual Automatic Traffic Light [5] performs its functioning through
Controlling, in which more man-power is required. Manually sensors and timers. Depending on timer values, lights are
it is not possible to control traffic effectively.. Another automatically getting ON and OFF. Another technique is
system is Automatic Controlling, which uses timer for each using sensors where wastage of time by constant on green
phase. Electronic sensors are used to detect vehicles and light on a bare road.. Hence, by using Automatic Traffic
produce signals. Time may get wasted by a green light on Light traffic congestion is still an issue.
empty road. All these limitations are supposed to be
eliminated via Image Processing, in which detection of 3. Literature Review
vehicles is done through pictures in place of installed sensors.
Firstly, traffic density is measured at signals and accordingly Traffic congestion at signals and roads is a serious issue.
change time delays for traffic lights the side at which the Ample of traffic light signals are there in India to control
traffic is more. In today’s era, Transportation is very traffic but still it is not efficient. Some traffic control system
important, because of which traffic is increased, and to are designed and implemented using a suitable
reduce this we need better traffic control system, which is Microcontroller [6], while some are implemented using
able to cope up with traffic situation and change controlling Embedded System.
as per the situation. Such an intelligent system will reduce
traffic without affecting transportation. Dynamic system will 3.1 Embedded systems
give a new look which can take dynamic decision like change
the time of red, green and amber lights. In this system, IR (infrared)[7] sensors, and built in 8
channels ADC are used. Emergency vehicles are detected
2. Existing Systems using programmable IR sensors. Drawback : IR sensors need
to be secured in safe place. Hence this system is expensive to
Various existing systems for traffic control are: implement.

Microcontroller[5] that controls the traffic lights at the zebra 3.2 Wireless Sensor Networks
crossing or we can say traffic junction, which is not a flexible
method, rather, there is fixed on and off timings for yellow, Fuzzy [7] logic is used to define direction of emergency
green and red lights. vehicle. Collection of information and response in an
effective way is done by central monitoring system.
Vehicle Actuated Control [5], continuously attempt to adjust Drawback: Communication in Wireless Sensor Network is
green light times. One of the major drawback is that it does still a research field. In order to survive in India’s weather,
not take into account the vehicles waiting at the red. A Sensors need to become robust.
detector is located at a distance ahead of stop line and it

Volume 5 Issue 8, August 2016

www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20161219 1396
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2013): 6.14 | Impact Factor (2015): 6.391
3.3 Active RFID and GSM Technology identifying points at which the image brightness changes
sharply, has discontinuities or noise. Value of brightness
System includes active RFID[7] tag, Wireless Coordinator, varies for different colours. Most bright colour is green.
Wireless Router, GSM modems and a monitoring station Image is blurred in blue, and image is noisy in red. Various
software. Data is collected by wireless devices, that are techniques are Sober Edge Detection, Pewitt Edge Detection,
mounted along the roadside, from Active RFID tags. Roberts Zerocoss Edge Detection and canny Edge Detection.
Monitoring station collects data through GSM and respond Comparing a reference image with the real time image pixel
accordingly. Drawback: Costly implementation of the system. by pixel is termed as Image Matching. Edge demarcates
between two different regions. Pixels are located in image
3.4 Image Processing which corresponds to boundaries of cars detected in the
image, which results to binary image of edge pixels detected.
Traffic scene is overlooked by cameras [7] at high posts of Sobert, Prewitt and Laplacian[5] operators are commonly
traffic lights. Extracted images from video are analyzed for used algorithms. In Gradient based Edge Detection[8], edges
detection and vehicles counting. After surveying we can are detected when we look in the first derivative for
reach to conclusion that Image processing is most efficient minimum and maximum values, where presence of edge is
technique among all the existing methods in terms of determined by first derivative and second derivative
efficiency, reliability, functionality, etc. determines where the pixel lies, either light or dark side.

Here we propose a system called Intelligent Traffic Control 3.5 Conclusion

[8] using Image Processing, in which, vehicles are detected
using cameras, which is placed along traffic light. An image Traffic management using image processing eliminates all
is a rectangular graphical object. Smallest element of an the shortcomings of earlier standard systems used for
image is a pixel; picture element, involves various issues controlling traffic. Excessive manpower is required in manual
ragarding to image compression, enhancement techniques controlling, while timer is used in automatic controlling had a
and various other operations which are complex. Such drawback of time being wasted by green light on a empty
operations are for enhancement of images such as road. Image processing eliminates all these pitfalls. This
sharpening, blurring, brightening, etc. It is kind signal technique is far more effective in traffic control. It diminishes
processing, in which, our image is the input signal. Output is the use of extra hardware devices like sensors, wireless
an image or parameters of an image. Mostly, image is treated routers, GSM modems, setup for a monitoring station etc.
as two-dimensional signal. Signal processing techniques are Presence of vehicles detection is consistent as we are using
applied to get the desired output. Image processing was actual images of traffic here. Reality is visualized and hence,
mainly digital, but now analog Image Processing also came functionality is more effective and efficient than all
into existence. Cameras are used to capture images of roads techniques. It achieves near perfect accuracy and
which refletcts traffic flow. Image Processing is beneficial performance of system is remarkable.
and economical as cameras are affordable than electronic
devices, sensors. We use a technique where captured image References
is compared to the referenced image, and various image
processing operations are applied to get the result. For image [1] A. Bonnaccorsi, “On the Relationship between Firm
morphology, finding optimal threshold value is very Size and Export Intensity,” Journal of International
important. Operations in the mathematical morphology are Business Studies, XXIII (4), pp. 605-635, 1992.
Erosion[9] and Dilation. They basically work for binary (journal style)
images and some of them also work for grayscale images. [2] R. Caves, Multinational Enterprise and Economic
An image is 2-D function as f(x,y), x, y are polar coordinates. Analysis, Cambridge University Press, Cambridge,
Amplitude is known as the intensity of the image. Conversion 1982. (book style)
of these y, x coordinates to digital values is required for [3] M. Clerc, “The Swarm and the Queen: Towards a
getting it processed through digital computer. At some stage, Deterministic and Adaptive Particle Swarm
image scaling occurs in all digital images. Image resizing is Optimization,” In Proceedings of the IEEE Congress on
necessary for increasing or decreasing the density of pixels. Evolutionary Computation (CEC), pp. 1951-1957,
Algorithm differs and hence, output of resizing image also 1999. (conference style)
differs. Every camera is resolution dependent, so system [4] H.H. Crokell, “Specialization and International
designed for one camera specification will not work for Competitiveness,” in Managing the Multinational
system designed for another camera specification. So, it is Subsidiary, H. Etemad and L. S, Sulude (eds.), Croom-
necessary to keep resolution constant, hence image resizing is Helm, London, 1986. (book chapter style)
performed. Perception of colours by human is through cones [5] K. Deb, S. Agrawal, A. Pratab, T. Meyarivan, “A Fast
which are sensitive to wavelength. One is green sensitive, Elitist Non-dominated Sorting Genetic Algorithms for
other is red sensitive, and the other is blue sensitive. We can Multiobjective Optimization: NSGA II,” KanGAL
obtain any colour from their combinations as we know these report 200001, Indian Institute of Technology, Kanpur,
are the three standard colours used. To convert R-G-B value India, 2000. (technical report style)
to gray, we know that each pixel has different R-G-B [6] J. Geralds, "Sega Ends Production of Dreamcast,"
parameter, consider each and obtain reflected single colour. vnunet.com, para. 2, Jan. 31, 2001. [Online]. Available:
Image enhancement is required for more appropriate results.
Edge detection is a set of mathematical methods which aim at
Volume 5 Issue 8, August 2016
www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20161219 1397
 International Journal of Science and Research (IJSR)
ISSN (Online): 2319-7064
Index Copernicus Value (2013): 6.14 | Impact Factor (2015): 6.391
http://nl1.vnunet.com/news/1116995. [Accessed: Sept.
12, 2004]. (General Internet site)

Author Profile
Parichita Basak is pursuing B.Tech in Information
Technology from Guru Gobind Singh Indraprastha
University, 2013-2017. She has completed Diploma in
Computer Science and Engineering from Board of
Technical Education in year 2014.

Ramandeep Kaur is pursuing B.Tech in Information

Technology from Guru Gobind Singh Indraprastha
University, 2013-2017. She has completed Diploma in
Computer Science and Engineering from Board of
Technical Education in year 2014.

Volume 5 Issue 8, August 2016

www.ijsr.net
Licensed Under Creative Commons Attribution CC BY
Paper ID: ART20161219 1398