IRJET- Artificial Intelligence Based Smart Traffic Management System using Video Processing

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 05 Issue: 03 | Mar-2018 www.irjet.net p-ISSN: 2395-0072
© 2018, IRJET | Impact Factor value: 6.171 | ISO 9001:2008 Certified Journal | Page 2271
Artificial Intelligence Based Smart Traffic Management System Using
Video Processing
Abhijeet Choudhary1, Akash Gupta2, Akshay Dhuri3, Prof. Nilima Nikam4
1,2,3 Final Year student, Department of Computer Engineering
4 Assistant Professor, Department of Computer Engineering
Y.T.I.E.T., Karjat, Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Due to increasing number of vehicles traffic
jams are becoming a common scenario in the whole country
as well as in the world. These frequent traffic jams at major
junctions kill a lot of man hours. Thus it creates a need for
an efficient traffic management system. So here we are
going to implement a smart traffic control system which is
based on the measurement of traffic density using real time
video processing technique. Through this paper we tried to
present a progress in the existing manual traffic control
system
Key Words: Gaussian mixture model, Shortest Job
First, Initialize Foreground Detector, Detect Cars in an
Initial Video Frame, Threshold, Traffic Density.
1. INTRODUCTION
As the population of the modern cities is increasing day by
day due to which vehicular travel is increasing which lead
to congestion problem. We used video processing
technique caused by using this we can easily calculated
density of traffic present on road. The system will detect
vehicles through images instead of using electronic
sensors embedded in the pavement. A camera will be
installed alongside the traffic light. It will capture image
sequences. Image processing is a better technique to
control the state change of the traffic light.
In our proposed system there we will be four cameras
in one intersection for a four way road. A CPU will be
connected with these cameras which will be responsible
for video processing. This processing unit take picture
from camera and compare all picture and take count the
vehicle present on the road. After comparison allocated
time first on that road where vehicle count in more, this
process happened again and again and reduced the traffic
conjunction.
Advantages
1. Heavy traffic jam reduced.
2. Decreased the pollution.
3. save human time which waste in traffic.
4. Save fuel and money.
1.1 Literature Survey
Traffic is a critical issue of transportation system in most
of all the cities of Countries. This is especially true for
countries where population is increasing at higher rate.
There is phenomenal growth in vehicle population in
recent years. As a result, many of the arterial roads and
intersections are operating over the capacity and average
journey speeds on some of the key roads in the central
areas are lower than 10 Km/h at the peak hour. In some of
the main challenges are management of more than
36,00,000 vehicles, annual growth of 7–10% in traffic,
roads operating at higher capacity ranging from 1 to 4,
travel speed less than 10 Km/h at some central areas in
peak hours. It involves a manual analysis of data by the
traffic management team to determine the traffic light
duration in each of the junction. It will communicate the
same to the local police officers for the necessary
actions.[1]
Reinforcement learning for traffic light control has
first been studied by Thorpe He used a traffic light-based
value function, and we used a car based one. Thorpe used
a neural network for the traffic-light based value function
which predicts the waiting time for all cars standing at the
junction. Furthermore, Thorpe used a somewhat other
form of RL, SARSA (State- Action, Reward-State Action)
with eligibility traces [2]. Roozemond describes an
intelligent agent architecture for traffic light control
intelligent traffic signaling agents (ITSAs) and Road
Segment Agents (RSAs) try to perform their own tasks,
and try to achieve local optimality. One or more Authority
Agents can communicate with groups of ITSAs and RSAs
for global performance. All agents act upon beliefs, desires,
and capabilities. No results were presented [3].
In G. Sathya, et al[3]achieved with the help of “AARS
using GPRS 3G TECHNOLOGY”. Through this, we can
provide a smooth flow for the ambulance by controlling
the traffic light according to the ambulance location to
reach the hospital. The location of the ambulance can be
easily identified with the help of the GPS unit installed in
it.[4] Then comes the Traffic light system using image
processing. The system will detect vehicles through
images instead of using electronic sensors embedded in
the pavement. A camera will be installed alongside the
traffic light. It will capture image sequences. [5].

1.2 Problem definition
All of the traffic light system used is the traditional system.
These systems encounter many limitations i.e. timing is
not based on number of vehicles due to this we have the
following drawbacks.
1. Heavy traffic jams.
2. Violation of traffic rules.
3. Wastage of man hours daily.
4. Increase in pollution in the consistent area.
5. Green Light for an empty road.
6. No traffic, but the pedestrians still need to wait.
7. Loss of Fuel and Money.
2. Proposed System
System is based on the measurement of traffic density
using real time video processing technique. The computed
traffic density is compared with other parts of the traffic in
order to control the traffic signal smartly. In this model,
there will be four cameras in one intersection for a four
way road. The hardware's that we will be using are: HD
Camera, CPU (For video processing), (we can install n
number of cameras to resolve the congestion problem for
n number of roads).
A High definition camera placed on poles will observe the
vehicular traffic flow continuously on a road then using
frame by frame Real-time video analysis through our
developed algorithm, we can detect how much cars are
present on the road.
Depending on the number of detected vehicles we have
developed and implemented a sequential traffic timer
system. Micro controller will detect the signal from CPU
and start the sequential traffic light. While the light phase
goes from green to red, our micro controller or Arduino
will send a signal to CPU.
The HD camera will be installed in the traffic light post at a
height of 19-25 feet (for practical implementation) above
the road. This camera will take the live video footage of
the road and send it to a computer where video analysis
will be done.
For a 4 way intersection, CPU will detect each and every
car and will count the vehicle number in the road by using
our developed algorithm. It will also do the same thing
with other road by using another camera.
CPU then compares vehicle number of both roads. The
road which has more vehicles will get the preference and
green light for that road will be on and red signal will be
shown automatically to the other road.
2.1 Algorithm
Gaussian mixture model is a probabilistic model for
representing the presence of sub-populations within an
overall population, without requiring that an observed
data set should identify the sub-population to which an
individual observation belongs.
A Gaussian mixture model is a distribution assembled
from weighted multivariate Gaussian* distributions.
Weighting factors assign each distribution different levels
of importance. The resulting model is a super-position (i.e.
an overlapping) of bell-shaped curves.
The Basic Formula for GMM is
Gaussian mixture models are semi-parametric. Parametric
implies that the model comes from a known distribution
(which is in this case, a set of normal distributions). It’s
semi-parametric because more components, possibly from
unknown distributions, can be added to the model.

Shortest Job First
Shortest job first (SJF) or shortest job next, is a scheduling
policy that selects the waiting process with the smallest
execution time to execute next. SJN is a non-preemptive
algorithm.
Shortest Job first has the advantage of having minimum
average waiting time among all scheduling algorithms.
It is a Greedy Algorithm.
It may cause starvation if shorter processes keep coming.
This problem can be solved using the concept of aging.
It is practically in feasible as Operating System may not
know burst time and therefore may not sort them. While it
is not possible to predict execution time, several methods
can be used to estimate the execution time for a job, such
as a weighted average of previous execution times. SJF can
be used in specialized environments where accurate
estimates of running time are available.
3. Implementation detail
Implementation of Smart Traffic system mainly consists of
2 parts. 1st part is image processing part and second is
signal controlling part. Front end of the system is Matlab.
Matlab will process all the video processing work and lest
controlling of the signal will be carried out by controller or
arduino.
3.1 Video Processing Using Matlab :-
Step 1 - Get Video and Initialize Foreground Detector
Rather than immediately processing the entire video,
the example starts by obtaining an initial video frame in
which the moving objects are segmented from the
background. This helps to gradually introduce the steps
used to process the video. The foreground detector
requires a certain number of video frames in order to
initialize the Gaussian mixture model. After the training,
the detector begins to output more reliable segmentation
results. The two figures below show one of the video
frames and the foreground mask computed by the
detector.
Fig -1: Original Image
Fig -2: Foreground Image

Step 2 - Detect Cars in an Initial Video Frame
The foreground segmentation process is not perfect and
often includes undesirable noise. The example uses
morphological opening to remove the noise and to fill gaps
in the detected objects.
Noise Removed Image
Fig -3: Clean Foreground
Next, we find bounding boxes of each connected
component corresponding to a moving car by using vision.
Blob Analysis object. To highlight the detected cars, we
draw green boxes around them.
The number of bounding boxes corresponds to the
number of cars found in the video frame.
Fig -4: Detected Cars
Step 3 - Process the Rest of Video Frames and send
count to controller.
In this step we will process the rest of the video and send
the car count to the controller.
Fig -5: Different Thresholds
Fig -6: GUI of the project
In matlab we make above GUI for project. In that we
make four windows of four different cameras video input.
There detect the every vehicle on road and increased the
number of count of vehicles. After some interval time it
will arranged in descending order and allocated signal for
that road on which more number of cars are present. This
process happened again and again and we reduced the
traffic.
Fig -7: Signal allocated on different roads
Fig -8: System Architecture

4. System Antiquation
4.1 Hardware:
PC & Cameras: A PC is used as a central device for various
image processing operations and Cameras to capture the
video to execute the project.
4.2 Software:
MATLAB: It is used in the entire processing for signal as
well as image processing.
5. CONCLUSION
Video detection technology became a new frontier in
case of vehicle tracking because of its dependability. Each
area needs to be exclusively programmed and the RFID
equipping and maintenance is somewhat costly. Unlike
any other system, our system confirms high accuracy and
we are confident about its success and feasibility.
However, further research and development in this
management system could bring that extra edge. So far
we’ve made this system to ease the traffic law
enforcement agencies. Knowing about the traffic pressure
of the adjacent node would make the system more
artificially intelligent. We hope these methods will be
adopted as soon as possible so that the limitations we are
experiencing with present method can be overcome.
6. ACKNOWLEDGEMENT
It gives us immense pleasure to express our deepest sense
of gratitude and sincere thanks to our highly respected
and esteemed guide Prof. Nilima Nikam, Department of
Computer Engineering Y.T.I.E.T., Karjat , for her valuable
guidance, encouragement and help for completing this
work. Their useful suggestions for this whole work and co-
operative behavior are sincerely acknowledged.
I also wish to express my gratitude to Prof. Harish
Barapatre (Computer Department) for his kind hearted
support. I also wish to express my indebtedness to my
parents as well as my family members whose blessings
and support always helped me to face the challenges
ahead. At the end I would like to express my sincere
thanks to all my friends and others who helped me directly
or indirectly during this project work.
7. REFERENCES
[1] Rajeshwari Sundar, Santhoshs Hebbar, and
Varaprasad Golla “Implementing Intelligent Traffic Control
System” Sensor Journal, IEEE(Volume:15, Issue :2)
[2] Ms. Sarika B. Kale and Prof. Gajanan P. Dhok
“Design of intelligent traffic light controller using
embedded system” Second International Conference on
Emerging Trends in Engineering and Technology, ICETET-
09.
[3] Xia and Shao, “Modelling of traffic flow and air
pollutions ion with application to Hong Kong Island”,
Journal of Environmental Modelling & Software. Vol. 20,
2005. pp 1175-1188.
[4] G.Sathya, Fathima Shameema S, Jyothi
MolSebastian, Jemsya K S“Automatic Rescue System for
Ambulance and Authoritative Vehicles, Vol.2 -Issue 4 April.
[5] Md. Rokebul Islam*, Nafis Ibn Shahid*, Dewan
Tanzim ul Karim*, Abdullah Al Mamun**, Dr. Md. Khalilur
Rh “An Efficient Algorithm for Detecting Traffic Congestion
and a Framework for Smart Traffic Control System”
Journal, IEEE Jan 20.

IRJET- Artificial Intelligence Based Smart Traffic Management System using Video Processing

More Related Content

What's hot (20)

Similar to IRJET- Artificial Intelligence Based Smart Traffic Management System using Video Processing (20)

More from IRJET Journal (20)

Recently uploaded (20)

IRJET- Artificial Intelligence Based Smart Traffic Management System using Video Processing