Traffic Light Controller System using Optical Flow Estimation

International Journal of Computer Applications Technology and Research
Volume 4– Issue 10, 755 - 761, 2015, ISSN: 2319–8656
www.ijcat.com 755
Traffic Light Controller System using Optical Flow
Estimation
Neha Parashar
Department of Computer Science and Engineeirng
Mewar University
Chittorgarh, India
Shiv Kumar
Department of Computer Science and Engineeirng
Mewar University
Chittorgarh, India
Abstract: As we seen everyday vehicle traffic increases day by day on road is causing many issues. We face many traffic jams due
to the inefficient traffic controlling system which is unable to cope up with the current scenario of traffic in our country. To overcome
such drastic scenario and looking at current traffic volume we need to develop a system which works on real time processing and works
after determining the traffic density and then calculating the best possibility in which the traffic on particular cross road is dissolved.
Also, it helps in saving time as on traffic roads. In present traffic control system when there is no traffic on road but the static signal not
allow traffic to move to cross and it changes after at fixed interval so at every cycle this amount of time is wasted for unused traffic
density road and if one road is at high traffic it continuously grows till human intervention. The basic theme is to control the traffic using
static cameras fixed on right side of the road along top of the traffic pole to check the complete traffic density on other side of the road.
This system will calculate number of vehicles on the road by moving detection and tracking system developed based on optical flow
estimation and green light counter will be based on the calculated number of vehicles on the road.
Keywords: Optical flow estimation, Moving object detection, tracking, Morphological operation, Blob analysis, Camera.
1. INTRODUCTION
In the current scenario of fixed time traffic lights, many a times
situation arises wherein there is heavy incoming traffic only
from one side of the intersection and the rest are relatively
empty. In this case the people on the heavily occupied side have
to wait for unreasonably long time as the green light timer is
fixed for each side which fails to take in to account that there is
no traffic to pass from the other sides. This prolonged waiting
time increases the average waiting time of every person in the
traffic. Though care is taken while setting these timers by
government officials according to the proportionate amount of
traffic present on different sides of the intersection but this can
never be so flexible as to adapt to the dynamic traffic
throughout the day. Moreover, some areas of high traffic
volume may receive scanty traffic at some point of the day and
some low traffic volume areas might get congested likewise,
which leads to an additional increase in waiting time because
these timers are set according to average volume of traffic
corresponding to the different areas of the city respectively but
not according to the different hours of the day. This fixed nature
of the present traffic light timers turns out to be ineffective not
only during the day but also during late night hours. At night,
when there is negligible traffic present, the timers lead to an
unreasonably long waiting time. Additionally most people,
seeing empty roads ahead of them, tend to jump lights at this
point, which quite often lead to accidents [1].
Our proposed system takes into account all these issues by
dynamically changing traffic light timers. It intelligently
recognizes the volume of the traffic at each side of the
intersection thereby providing an adequate amount of time for
the traffic to pass. The system monitors traffic throughout the
day and takes care of the inability of the fixed timers to adjust

www.ijcat.com 756
as per the traffic. Since our system regulates the flow of traffic
at night also, it helps in minimizing the chances of an accident.
2. LITERATURE SURVEY
E. Atkociunas, R. Blake, A. Juozapavicius, M. Kazimianec
(2005) presented paper on “Image processing in road traffic
analysis”. In this Researches and developments have been
performed in image a processing technique which is applied to
traffic data collection and analysis [2].
Pejman Niksaz (2012) has proposed a paper “Automatic traffic
estimation using image Processing” that estimates the size of
traffic in highways by using image processing has been
proposed and as a result a message is shown to inform the
number of cars in highway [3].
P.D. Kamble, S.P. Untawale and S.B. Sahare (2012) presented
paper on “Application of image processing for traffic queue
length” to measure queue parameters accurately, traffic queue
algorithm is used [4].
Raad Ahmed Hadi, Ghazali Sulong and Loay Edwar George
(2014) presented paper on “Vehicle Detection and Tracking
Techniques: A Concise Review” in which they present a
concise overview of image processing methods and analysis
tools which used in building these previous mentioned
applications that involved developing traffic surveillance
systems [5].
Ramesh Navi, Aruna M. G. (2014) presented paper on “Traffic
Event Detection using Computer Vision” they proposed a
system in which process includes the subtasks of data
collection, negative positive separation, creating training
samples, creating description files, haar training and a strong
action program which can detect vehicles and the traffic events
[6].
Pallavi Choudekar, Sayanti Banerjee, Prof. M.K.Muju
presented paper on “Real Time Traffic Control using Image
Processing” propose a system for controlling the traffic light by
image processing. The system will detect vehicles through
images instead of using electronic sensors embedded in the
pavement [7].
Overall, the references have advantages such as reporting of
speed violation, traffic congestion, accidents, low cost and
setup with good accuracy and speed. Some of the disadvantages
occurring are variation of ambient light, 3D images are not
supportive, and it is difficult to detect vehicle features in windy
and other weather conditions.
3. PROBLEM STATEMENT
Most of Traffic Controllers in India are Manual Controlling and
Fixed Time Traffic Signalization. They lead to traffic accidents
at intersections. These crashes are caused by drivers’
frustration because of long intersection delays. Without police
or other forms of enforcement, the long delays have led to road
crashes caused by drivers’ disobedience of traffic signals and
dangerous driving maneuvers in an effort to beat the signal so
as to avoid the long delay. This is owing to the fact that some
drivers will not wait for the green times when there are no
vehicles approaching the intersection, thus Red Light Running
(RLR) will occur. In addition, drivers will try to change lane so
as to be close to the stop line. The fuel emission and
environmental aspects have an influence on people living near
the intersections.
4. OBJECTIVE
1. To minimize estimated vehicle delays and stops.
2. To maximize intersection capacity.
3. To design automate control system for traffic on
streets.
5. PROPOSED SYSTEM
A scene should be selected from a static camera. Our test
movies are selected from urban surveillance videos. Some pre-
processing operations have to be done for making the scene
ready to process. After that propose algorithm apply on video
show below:
RGB to Intensity
Mean Calculation
Optical Flow Estimation
Median Filter
Morphological Close
Video Frame Acquisition

www.ijcat.com 757
Rules for green light signaling & time control:
Rule 1: If
Number of vehicle < 0
Then green light counter = 5 sec
Rule 2: If
Rule 3: If
Else
green light counter = 60 sec
In proposed system we set maximum time counter to 60 sec and
minimum time counter to 5 sec. Here we show three type of
case. We decide the green light counter according to distance
and time. In how much time a particular distance travel by
vehicle. The counter time depends on distance because the
distance at every intersection road is different.
6. THEORETICAL WORK
To design advance traffic control system firstly we setup a
camera at 11 feet height. Record video from the camera and
perform multiple operations on it discussed below:
6.1 RGB to Intensity
First step for this project is to convert the video we intend
to simulate from RGB to intensity. RGB (red, green, blue)
are the three colors that can be mixed to become any other
colors in the color spectrum. In an RGB video, each pixel
is represented by a combination of these colors. Though it
provides a more accurate visual representation of the
recorded object(s), having to detect 3 colors in every pixel
is redundant. Hence, the simplest way is to convert the
video from RGB to intensity. What converting the video
to intensity does is represent each pixel in the video with
a value ranging from 0 to 255. 0 being the color black; 255
being the color white. Any values in-between are shades
of gray.
6.2 Mean Calculation
The mean of every frame is calculated on gray-scale
format.
6.3 Optical Flow Estimation
Optical flow has been used in conventional video
surveillance systems to detect motion, but the purpose of
using optical flow in such systems is just to detect moving
objects. However, the purpose of using optical flow in our
system is to provide statistical traffic flow information.
The role of optical flow in our system is essential and
critical to final performance. Thus, application-specific
assumptions must be considered to choose an algorithm
for optical flow estimation from among the many
available optical flow algorithms [10].
The optical flow method tries to calculate the motion
between two image frames which are taken at times t and
t+δt at every position. These methods are called
differential since they are based on local Taylor Series
approximation of the image signal; that is, they use partial
derivatives with respect to the spatial and temporal
coordinates.
There are two most used methods, namely:
 Lucas-Kanade
 Horn-Schunck
To solve the optical flow constraint equation for u and v,
the Lucas-Kanade method divides the original image into
smaller sections and assumes a constant velocity in each
section. Then, it performs a weighted least-square fit of
the optical flow constraint equation to a constant model
for in each section, Ω, by minimizing the following
equation:
Blob Analysis
Tracked Detected Objects
Count Number of Vehicle
Traffic Light Control
(Control on the basis of number of vehicle)

www.ijcat.com 758
6.4 Median Filter
In signal processing, it is often desirable to be able to
perform some kind of noise reduction on an image or
signal. The median filter is a nonlinear digital filtering
technique, often used to remove noise. Median filtering is
very widely used in digital image processing because,
under certain conditions, it preserves edges while
removing noise [11].
6.5 Morphological Close
Morphological operations are performed to extract
significant features from images that are useful in the
representation and description of the shapes in the region;
mostly used in image segmentation and pattern
recognition. In the proposed system we used both
morphological close and erode, respectively, to remove
portions of the road and unwanted objects. After
morphological closing operation, on condition that
vehicle’s appearance is not destroyed, objects including
many small holes and separated pixels may be connect
into one big actual vehicle shape.
6.6 Blob Analysis
In the area of computer vision, blob detection refers to
visual modules that are aimed at detecting points and/or
regions in the image that differ in properties like
brightness or color compared to the surrounding.
6.7 Vehicle Detecting, Tracking and
Counting
Blob analysis has the functionality to produce many forms
of statistics, which is crucial for detecting and tracking.
For now, the bounding box option is checked. A bounding
box is an M-by-4 matrix of [x y height width] bounding
box coordinates, where M represents the number of blobs
and [x y] represents the upper left corner of the bounding
box. As the blob moves, the bounding box will follow.
Tracking vehicles with boundary boxes are counted.
6.8 Calculation for Green Light Timer of
Traffic Light Control System
In this section proposed system provide counter time for
green signal light according to the number of the vehicles
are present in the video. The counter time for green light
decide according to the distance between the intersection
and the speed of the vehicle.
In proposed system we manage only green light counter only.
System, calculate the number of vehicle present in video.
According to the number of vehicle presented in the video set
green light timer. The timer for green light varies as number of
vehicle varies in detected video. Green light time predefined in
the system for different-different condition of the counted
vehicle. According to perfect match condition for detected
vehicle found system send the green light counter time.
To decide the different-different green light timer depends
upon some parameters. These parameters are: distance and
speed. Distance parameter, depends upon the distance which is
travelled by the vehicle to cross the intersection area. Distance
is different-different for different intersection area. Speed,
speed of the waiting vehicles is different from each other and
everybody drive their vehicle on different speed. To calculate
time of green light signal use a basic formula.
Time = Distance/Speed
Time = sec
Distance = meter
Speed = meter/sec
In our proposed system we calculate time on the basis of the
parameter which is measured by us to calculate signal light
time. In this system we calculate and measure parameter from
the intersection road at Bhilwara. At this intersection road/area
the 100 meter distance travelled by vehicle within 10 sec. by
average speed. So we take the distance as 100 meter. Speed for
the system we take 30 km/hr. This is calculated on the basis of
average speed of all vehicles. According to Indian govt. speed
limit for all vehicle is decided. On the basis of their speed we
take the average speed.
According to given formula time decided for proposed system:
Time =
100 meter
30 km/hr
Time = 11 sec.
A vehicle which is within the 100 meter (including intersection
road/area) road is easily able to cross the road within 10 sec.
7. RESULT
In this section we show the experimental result using data from
traffic video. Before applying optical flow estimation on frame
the image RGB converted into intensity (gray). Original video
and after intensity and median filter shown in fig 1

www.ijcat.com 759
(a)
(b)
(c)
(d)
Figure 1: (a) Original Video (b) Intensity (c) Before filter (d)
Median filter
The video analyzed for detecting and counting number of
vehicle within the video frame. According to number of vehicle
it had shown the time period for green signal light. The system
was able to produce results for different scenarios.
Case 1: If Number of vehicles is less than 15 then green signal
light timer set to 10 sec.
Figure 2: No. of vehicle = 8, Green light time = 10 sec
Case 2: If Number of vehicles is less than 30 then green signal
light timer set to 30 sec.

www.ijcat.com 760
Figure 3: No. of vehicle = 18, Green light time = 30 sec
Case 3: If Number of vehicles is more than 30 then green
signal light timer set to 60 sec.
Figure 4: No. of vehicle = 35 Green light time = 60 sec
As experimental result shows green light time counter decided
according to number of vehicle detected in the video. When the
number of vehicles are increase at signal light green light
counter also vary.
8. RESULT ANALYSIS
In this chapter we present the test environment and the
experimental results analysis of our system.
No of
vehicle
Present
System timer
(sec.)
Proposed
System
timer (sec.)
Saved
Time
(sec.)
0 60 5 55
<15 60 10 50
<35 60 30 30
>35 60 60 0
Table 1: Result Analysis
In graphical format result analysis can be shown as given
below:
Figure 5: Result Analysis
In present system fuel consumes a lot and it increase air
pollution also. According to result analysis it shows we can
save a lot of time at traffic signal. Lot of fuel consume by
vehicles in waiting time can also be saved. Our proposed
system also helps in decrease the fuel consumption by vehicle
and decrease the import of fuel from the other countries and
saving the foreign currency.
9. CONCLUSION
In this thesis we showed that Control Traffic Light Controller
using Optical Flow Estimation. We demonstrate on traffic
density estimation and manage traffic signal light based on
image processing technique and we successfully calculated the
traffic and also manage green light counter. It can be reduce
the traffic congestion and time wasted by a green light on an
empty road. It shows that it can reduce the traffic congestion
and avoids the time being wasted by a green light on an empty
road. It is also more consistent in detecting vehicle presence
because it uses actual traffic images.
10. FUTURE WORK
Future work will cover complex testing of the system, and more
detailed development of modified algorithms. Also the
0
10
20
30
40
50
60
70
0 <15 <35 >35
Present System
timer (sec.)
Proposed
System timer
(sec.)
Saved Time
(sec.)

www.ijcat.com 761
algorithm and identify overlapping objects transparencies
during object tracking.
11. ACKNOWLEDGMENTS
I would like to thanks, Mr. Shiv Kumar and all CSE department
of Mewar University. I also want to thanks my family.
12. REFERENCES
[1.] Pranav Maheshwar, Deepanshu Suneja, Praneet Singh,
Yogeshwar Mutneja, “Traffic Image Processing
System”, Volume 118 – No.2 3, May 2015
[2.] E. Atkociunas, R. Blake, A. Juozapavicius,M.
Kazimianec “Image processing in road traffic analysis”
Nonlinear Analysis: Modelling and Control, 2005, Vol.
10, No. 4, 315–332.
[3.] Pejman Niksaz, “Automatic traffic estimation using
image Processing” International Journal of Signal
Processing, Image Processing and Pattern Recognition
Vol. 5, No. 4, December, 2012.
[4.] P.D. Kamble, S.P. Untawale and S.B. Sahare,
“Application of image processing for traffic queue
length” VSRD-MAP, Vol. 2 (5), 2012, 196-205.
[5.] Raad Ahmed Hadi, Ghazali Sulong and Loay Edwar
George, “Vehicle Detection and Tracking Techniques:
A Concise Review”, Vol.5, No.1, February 2014
[6.] Ramesh Navi, Aruna M. G., “Traffic Event Detection
using Computer Vision”, Volume 16, Issue 3, Ver. II ,
May-Jun. 2014.
[7.] Pallavi Choudekar, Sayanti Banerjee, “Real Time
Traffic Control using Image Processing”, Vol. 2 No. 1.
[8.] Joonsoo Lee and Alan C. Bovik, “Estimate and
Analysis of urban Traffic flow”, 2009.
[9.] Pre Kumar.V, Barath. V, Prashanth.K, “Object
Counting and density Calculation using MATLAB”.

Traffic Light Controller System using Optical Flow Estimation

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to Traffic Light Controller System using Optical Flow Estimation (20)

More from Editor IJCATR (20)

Recently uploaded (20)

Traffic Light Controller System using Optical Flow Estimation