Vehicle Traffic Analysis using CNN Algorithm

International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 01 | Jan 2022 www.irjet.net p-ISSN: 2395-0072
© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 1375
Vehicle Traffic Analysis using CNN Algorithm
Pragati Bhosale1, Ankita Kawatikawar2, Pritee Jadhav3, Prof.Sonali Patil4
1Student, Dept. of Information Technology, Dr.D.Y.Patil Institute of Technology, Pimpri Pune, Maharashtra, India
2Student, Dept. of Information Technology, Dr.D.Y.Patil Institute of Technology, Pimpri Pune, Maharashtra India
3Student, Dept. of Information Technloogy, Dr.D.Y.Patil Institute of Technology, Pimpri Pune, Maharashtra, India
4Professor, Project Guide, Dept. of Information Technology, Dr.D.Y.Patil Institute of Technology, Pimpri Pune,
Maharashtra, India
Abstract - The goal is to build a traffic light system that
changes based on how many people are in the area. When
there is a lot of traffic at an intersection, the signal time
automatically changes. Many major cities around the world
have a lot of traffic, which makes it hard to get to work every
day. Traditional traffic signal systems are based on the idea
that each side of the intersection has a set amount of time.
They can't be changed to account for moretraffic. People can't
change the times of the intersections that have been setup for
them. There may be more traffic on one intersection, which
could make it more difficult for the typical greenperiodtoend.
After processing and translating the traffic signal object
detection into a simulator, a threshold is set anda contour is
drawn many cars are in the area. After , we can figure out
which side has the most carsbased on the signals sent to each
side. Paper provides a solution based on camera feed at
crossing for each lane process the data through and allocates
the ”green” time according to its traffic flow density using
YOLO v3 and also takes care of starvation issue that might
arise of the solution. As a result ,the flow of traffic oneachlane
is automatically optimized and the congestion that used to
happen unnecessarily is eliminated earlier and results show
significant benefits in reducing traffic waiting time
Key Words: CNN, Classification, Deep learning, Traffic
Analysis , traffic signal, deep learning, Congestion
detection ,YOLO v3 etc.
1. INTRODUCTION
Traffic control and management are essential issues in a
number of regions, particularly those with expanding
populations and large cities. Traffic lights utilize time
division multiplexing toalleviatecongestionatintersections.
Invarious countries,fixed-cyclecontrollersare employed
at all signalized intersections.The soledisadvantageofusing
a traffic light is the delay in reaching your destination (stop
time or waiting time). The delay at an intersection is a
performance indicator of a traffic signal controller's
efficiency. The phases, sequence,andtimingoftrafficsignals
all contribute to theefficiency of traffic movement across an
intersection. The adaptive signal controller is in charge
phases, sequence, and timing. When it comes to reducing
traffic congestion,the timing and sequence of traffic signals
must be optimized. Traffic signal time management is tough
and blind due to unpredictability and a plethora of other
factors. this project is to develop a real-time adaptive of
traffic signals.
The current traffic control system works based on time to
switch the traffic lights. But many researches are conducted
to change the current traffic light system into automatic and
adaptive system to solve the problems with the traffic
congestion. Some researchers used hardware installation
such as sensors and Radio Frequency Identification [8] to
detect the crowdedness of vehicles, but this is expensiveand
difficult to implement. Some researchers are alsoworkingto
solve the problem with the help of image processing using
image subtraction methodtocalculatethedensityofvehicles
[1] - [4]. They have used a fixed image that cannot be
changed, as a reference image in image subtraction method.
But this method is not efficient in the night-time,becausethe
light condition in the night-time is not same as in daytime.
The decision making for switching the traffic light works
based on the calculated density. Anurag [1] used an
algorithm to determine the approximate density of vehicles
on the road with fourlanes.Usingthisalgorithmthedynamic
system [1] improves 35% over the hard coded system.
Ashwini [2] used a motion detection algorithm to estimate
the count of vehicles on the road; the estimated count will
then used to control the traffic signal.
---------------------------------------------------------------------***---------------------------------------------------------------------
Although the importance of traffic lights which give safety
to the users on roads, the traffic jam causes great loose in
time and energy (fuel) for some people, while others
crossing road or roundabout have no traffic jam. The main
objective of this paper is to design and implement a
suitable algorithm and its simulation for an intelligent
traffic signal simulator. The system developed is able to
sense the presence or absence of vehicles within a certain
range by setting the appropriate duration for the traffic
signals to react accordingly. By employing mathematical
functions to calculate the appropriate timing for the green
signal to illuminate, the system can help to solve the
problem of traffic congestion. The reason depends on
resent fixed programming time. So, our target in this paper
is to make this time unfixed according to the size of traffic
jam, When there is a traffic jam in any road the green light
which means permeation gives full time to the user of the
road. If there is no traffic jam, the green light does not give

The proposed system is to develop a smart traffic light
switching with the techniques of image processing that can
switch the traffic signals in different ways for day-time and
night-time. In the day-time the system measures the density
of the vehicles on the road and in the night-time the system
counts the number of vehicles on the roadusingthevehicle’s
headlight, based on these measurements the traffic lightwill
be switched. Apart from that the proposed system will
improve the functionalities of the previous works, such that
it can detect the traffic violations such as a red light
violation, stop line and lane violations. Each of the lights will
have their own additional features, such that the red light
detects a stop line and red light violations, and the green
light will also detect the lane violation. In the proposed
system some filtering techniques, image enhancement and
segmentation will be used to remove a noise and improve
the quality of the captured image so that the accuracy and
efficiency of the system will be improved accordingly[11].
2. LITERATURE SURVEY
Vehicle Classification techniques Comparison by Machine
learning on roadside sensors shows thatThedatasetof3074
samples is processed for vehicle classification by using
different algorithms of machine learning. Various
classification techniques are used such as SVM, neural
networks and logical regression. Logical regression shows
the results had high performance when comparing with
other methods of machine learning with the classification
rate is 93.4% The main difficulty in this method is the usage
of datasets, as it was focused mainly on single class which is
very difficult to search while classification[13].
Comparison of vehicle type: Various Schemes of
Classification shows that Vehicles are classified into four
different classes car, bus, van and motorcycle. Two types of
methods used here, SVM and random forest which is a
feature. The accuracy of SVM is 96.26% morerobustthan RF
Due to similar image size and shape of car, bus and van,
miscalculation occurs[14].
vehicle detection and classification in real time video
streams Distributed method of real time vehicle detection
and classification system is proposed by Kul etal.[17].Other
techniques used here are vehicle classification, feature
extraction, detection of foreground and background
subtraction. In broad daylight the resultsarepromisingwith
an accuracy of 89.4% In night and bad conditions of weather
they didn’t perform any work. [16] Z. Dong, Y. Wu, M. Pei,
and Y. Jia Semi-supervised Convolutional Neural Network is
used for vehicle classification [15].
Semi-supervised Convolutional Neural Network is used
while the classification of vehicles. The dataset consists of
9850 high resolution images areused. Thedatasetholdsonly
front views of vehicles. In daylight 96.1% accuracy is
registered and in Night89.4%Misclassificationoccursdueto
incorrect labels in the BIT dataset[18].
Feature-Based Tracking The proposed method is feature
based tracking method which usesfeaturedescriptorofSIFT
for tracking. It forms a rich representation of object classes.
The proposed approach provides better performance.When
the view is changed the system is ineffective and occlusion
also not tested[19].
Color and Pattern Based Tracking The color and pattern of
vehicle image series of traffic video surveillance areusedfor
tracking. It consists of segmentation of foreground and
background, vehicle flow, shade removal, vehicle velocity,
vehicle count, vehicle locationtotrack objects.Thissystemis
proved to work in different climatic conditions and is
insensitive to lighting conditions The system needs to be
tested under extreme weather conditions and occlusion
problems also need to be checked[21].
3. PROPOSED SYSTEM
In this system we are taking input as an image. As weknow
that we are performing image processing operation on
system, so that we are using four modules of image
processing like preprocessing, segmentation, feature
extraction and classification where we use our CNN
algorithm. So first we have passed input as an image then in
preprocessing RGBconversionandthenBinaryconversionis
done then.
Fig.1. System Architecture
full time, but it gives programming time. The new timing
scheme that was implemented promises an improvement
in the current traffic light system and this system is
feasible, affordable and ready to be implemented
especially during peak hours[35].

After that in the segmentation part the image is divided into
the small pixels then after segmentation in the extraction
part system extract the geometry based feature of traffic
sign. then in classification where we use our CNN algorithm
to classify and prediction[8] ,we pass this geometry based
features oftrafficsigntotheclassificationtoforclassification
and prediction , then on that basis it detect the traffic signal
then convert it into the voice alert.
4. DESIGN AND METHODOLOGY
Fig.2 Workflow of proposed system
The system will use image extractionmethodtocalculatethe
amount pixels occupied by vehicles on the road. The
proposed system uses two different methods i.e. inday-time
and night-time. At day time, Density of vehicles will be
calculated , because the rate of vehicles are more visible in
the daytime than in the night time. So it is effective to use
density count instead of vehicle count in day-time. Counting
the number of vehicles in the daytime may lead to a false or
ambiguous result because two very close vehicles may be
counted as a one vehicle.
The proposed algorithm checks the time, if it is a day or
night in order to switch the system signal accordingly. The
decision module receives densitycount(numberofvehicles)
in green signal and red signals (2) (3).Basedonthesevalues,
the decision module will calculate the amount of the green
signal time (TDi and TNi) and decide which side of the road
will be switch to a green signal.
5. ALGORITHMS
5.1CNN (CONVOLUTION NEURAL NETWORK)
Computer vision and pattern recognition benefit greatly
from the use of fully convolutional networks. CNNs are
frequently employed in image analysis tasks such asimage
recognition, object recognition, and image segmentation.
Deep neural networks consist of four layers. In
traditional neural networks, each input neuron hiddenunit.
EachinputneuronLayerisonlylinkedtootherinputneuron
units. Only a few of CNN communicate with layerbe low it.
It's reducing the three-dimensionality the CNN's hidden
layer, activation and maximum pooling.Aone-dimensional
array is created by flattening data before moving is
generated by flattening Connected Tiers are the last few
nodes that are all linkedtogether completely. Fully linked
layers receive as input smoothedoutputfromprior pooling
or pooling layers. Sothat's how it works, as it were.
CNN implementation steps :
• Step 1: Convolution Operation(Filter image)
• Step 1(b): ReLU Layer
• Step 2: Pooling (used max pooling function)
• Step 3: Flattening (Covert Matrix into 1DArray)
• Step 4: Full Connection.
• Step 4(b): Dense()
• Step 4(c): Optimizer()
• Step 4(d) : compile()
5.2 YOLO V3(OBJECT DETECTION AND
CLASSIFICATION ALGORITHM)
YOLOv3 (You Only Look Once, Version 3) is a object
detection algorithm that perform real time object detection
that and detect the objects specified in image ,video and live
feeds. It uses features learned by a deep learning technique
based on convolutional neural network for object detection.
Implementation of YOLO is done by using Keras or OpenCV
deep learning libraries.

Fig.3 Flowchart of YOLO
5.3 VEHICLE DENSITY COUNT
The following are steps to calculate the density of
vehicles.
 Image acquisition: The proposed system will start
by capturing a live real time video, input video or
images using a video camera.
 Initially, the images capture by system will be
empty roads with no vehicles and itwill beusedasa
reference image RI. The system will capture a
continuous sequence image frames from the live
video or from given video/image per one second,
which is used as a current image (CI). For both
cropped for both referenceandcurrentimages Only
the interested target area of the road will be, to
eliminate the unnecessary parts.
 To separate the foreground objects (vehicles) from
the background the Background subtractionwill be
applied in each sequence of image frames, then the
result image (I) will be obtained .Processing of
subtracted image will be done by converting from
RGB (Red Green Blue) to Grayscale for further
processing.
 In each step of the image acquisition process, a
noise may be there so Image filtering techniques
will be applied to remove noises, here median
filtering will be used to remove pepper noises and
salt and will produce a filtered image.
 In the filtered generated result, image there maybe
some non vehicles detected as foreground .In order
to improve quality of result image the non vehicles
object need to be remove.. So that thresholding that
will be applied to differentiate the objects (white)
and non object (black). Dilation morphological
technique will also be used to fill the holes inside
vehicle objects; For examining and expanding the
shapes of the image and to extend the border and
regions of the objects dilation is used .
 This results the final black and whiteimage(Ibw).It
is further is used for calculation of density count.
Here if the pixel value [pv] is not a zero, which will
be considered as an object or vehicle
. But if pv is zero, which is considered as a
background (non object) that needs to be
eliminated. Ibw = 1 if pv ≥ 1 0 else (1) .
 Finally the density of vehicles on the road will be
calculated (not number of vehicles). The value of
vehicle density determines the amount for which
portion of the road is occupied by vehicles [4]. 𝐷=𝑛
𝑖=1 Ibw 𝑚 𝑗=1 (2) Here n is number of rows and m
is number of columns. Only the white pixel valuesin
all rows and columns will be added to density (D).
5.4 HOW THE SIGNAL WILL BE SWITCHED
 The density /count for the vehicles from sides of the
road is determined and will be used as a input
parameter to switch the signals.
 Green signal Timeiscalculatedusingdensityorcount
of vehicles in one road per the total density (vehicle
count) in all sides of the intersection road.
 The proposed method uses the formula in [4] to
calculate the green signal time, It will produce three
outputs from the input parameters given ; weighted
time(WD, WN) and trafficcycle(Tc). Totalamountof
time for one complete cycle of the traffic lights is
given by Tc.
 WDi is a weight factor at a particular road in the
intersection road will calculated as: WDi = Di n j=1 Dj
(4)
 WNi is a weight factor at a particular road in the
intersection road and will as: WNi = Ci n j=1 Cj (5)
Where WDi is a weight factor of ith road in day-time,
WNi is a weight factor of ith road at night-time,
density calculated in day-time is D, vehicle count
calculated in night-time is C, and the total number of
road in the intersection is N.
 The time (TDi) of green light at ith road in the day-
time is calculated by: TDi = Tc × WDi (6)
 The time(TNi) forgreen light that will be assigned to
ith road in the night-time is calculated by: TNi = Tc ×
WNi (7)
 Finally, this received value will be sent to signal
controller and it will switch the signals accordingly
based on the decision phase module. The maximum
green light provided to a lane must be 60 Sec and
minimum is 15 sec.
6. CONCLUSION
In order to record real-time traffic condition
notifications, we may integrate our system with an app
that analyses official traffic signals. As a result, in the
worst-case situation, our system will be able to signal
traffic-related events at the same time the console's
results are displayed on the websites.In termsoffeature
coverage, weare also investigating the integrationofour
system into a more extensive traffic monitoring

infrastructure. This infrastructure could include
improved physical sensors as well as social sensors like
social media streams. Social sensors, in particular, have
the potential to provide low-cost comprehensive
coverage of the roadnetwork, especially in areas where
traditional traffic sensors are sparse (e.g., urban and
suburban areas).The proposed strategy limits traffic
delay which helps to reduce traffic congestions,
environmental effects. The constraint of this work is,the
proposed technique relies upon the vision framework
introduced at convergence focuses that have variable
video properties. In a future work, try to install the
proposed method .
ACKNOWLEDGEMENT
We are thankful to our guide Prof. Sonali Patil who
provide us guidance ,support and expertise.Wearealso
thankful to our Principal Dr.Pramod Patil and HOD of
dept. of Information Technology Prof.S .A Nalawade for
the support .
REFERENCES
1) Anurag Kanungo, Ayush Sharma, Chetan
Singla, Smart Traffic Lights Switching and
Traffic Density Calculation using Video
Processing, Proceedings of 2014 RAECS
UIET Panjab UniversityChandigarh,06–08
March, 2014, 978-1-4799-2291-
8/14/$31.00 ©2014 IEEE
2) Ashwini D. Bharade, Surabhi S. Gaopande,
Robust and Adaptive Traffic Surveillance
System for Urban Intersections on
Embedded Platform, 2014 Annual IEEE
India Conference (INDICON), 978-1-4799-
5364-6/14/$31.00 ©2014 IEEE
3) M. Ashwin, B.K Arvind, R. Barath Kumar, S.
Arun Karthik, Pixel Detection and
Elimination Algorithm to Control Traffic
Congestion Aided by Fuzzy Logic, 2013
Fifth InternationalConferenceonAdvanced
Computing (ICoAC), 978-1-4799-3448-
5/13/$31.00 ©2013 IEEE
4) Md. Munir Hasan, Gobinda Saha, Aminul
Hoque, Md. Badruddoja Majumder, Smart
Traffic Control System with Application of
Image Processing Techniques, 3rd
international conference on informatics,
electronics & vision 2014, 978-1-4799-
6711-7/14 $31.00 © 2014 IEEE
5) Adi Nurhadiyatna, Wisnu Jatmiko, Benny
Hardjono Ari Wibisono1, Ibnu Sina, Petrus
Mursanto, Background Subtraction Using
Gaussian Mixture Model Enhanced by Hole
Filling Algorithm (GMMHF), 2013 IEEE
International Conference on Systems, Man,
and Cybernetics, 978-1-4799- 0652-9/13
$31.00 © 2013 IEEE
6) Marcos Paulo Batista, Patrick Y. Shinzato,
Denis F. Wolf and Diego Gomes, Lane
Detectionand EstimationusingPerspective
Image, 2014 Joint Conference on Robotics:
SBR-LARS Robotics Symposium and
Robocontrol. 978-1-4799-5180-
2/14/$31.00 ©2014 IEEE
7) Ramesh Marikhu, Jednipat Moonrinta,
MongkolEkpanyapongandMatthewDailey,
Supakorn Siddhichai, Police Eyes: Real
World Automated Detection of Traffic
Violations, 978-1-4799-0545-4/13/$31.00
c 2013 IEEE
8) Harpal Singh, Satinder Jeet Singh, Ravinder
Pal Si, Red Light Violation Detection Using
RFID, Proceedings of ‘I-Society 2012’ at
GKU, Talwandi Sabo Bathinda (Punjab)
9) Chandrasekhar. M, Saikrishna. C,
Chakradhar. B, Phaneendra Kumar. P &
Sasanka. C, Traffic Control using Digital
Image Processing, ISSN(Print):2278-8948,
Volume-2, Issue-5, 2013
10) A.H.M Almawgani “ Design of real time
Smarttrafficcontrolsystem‘‘Departmentof
Electrical Engineering, College of
Engineering, Najran University, Najran,
Saudi Arabia
11) Dipti Kapoor Sarmah “Smart Traffic Light
Controlling And ViolationDetectionSystem
Using Digital Image Processing”
https://www.researchgate.net/publication
/304131240 august,2016
12) Dr. S.V. Viraktamath Madhuri Yavagal
Rachita Byahatti “Object Detection and
Classification using YOLOv3
“http://www.ijert.org ISSN: 2278-0181
IJERTV10IS020078 Vol. 10 Issue 02,
February-2021.
13) D. Kleyko, R. Hostettler, W. Birk, E. Osipov,
"Comparison of Machine Learning
Techniques for Vehicle Classification Using
Road Side Sensors", 2015 IEEE 18th Int.
Conf. Intell. Transp. Syst., pp. 572-577,
2015.
14) Z. Chen, T. Ellis and S. A. Velastin "Vehicle
type categorization: A comparison of
classification schemes", 14th IEEE Annual
Conference on Intelligent Transportation
Systems, the George Washington
University, Washington,DC,USA.pp.74-79,
Oct. 5-7, 2011.

15) S. Kul, S. Eken, and A. Sayar, ―Distributed
andcollaborativerealtimevehicledetection
and classification over the video streams,‖
Int. J. Adv. Robot. Syst., vol. 14, no. 4, p.
172988141772078, Jul. 2017.
16) Z. Dong, Y. Wu, M. Pei, and Y. Jia, ―Vehicle
TypeClassification Using a Semisupervised
Convolutional Neural Network,‖ IEEE
Trans. Intell. Transp. Syst., vol. 16, no. 4, pp.
2247–2256, Aug. 2015
17) Girisha, R. and Murali, S. (2011). Tracking
Humans using Novel Optical Flow
Algorithm for Surveillance Videos,
Proceedings of the Fourth Annual ACM
Bangalore Conference, ACM, pp: 7.
18) Tzagkarakis, G., Charalampidis, P.,
Tsagkatakis, G., Starck, J.-L. and Tsakalides,
P. (2012). Compressive VideoClassification
for Decision Systems with Limited
Resources, Picture Coding Symposium
(PCS), 2012, IEEE, pp. 353–356.
19) M. Xiaoxu and W. E. L. Grimson, "Edge-
based rich representation for vehicle
classification," in Computer Vision, 2005.
ICCV 2005. Tenth IEEE International
Conference on, 2005, pp. 1185-1192 Vol. 2.
20) W. Hsieh, et al., "Automatic traffic
surveillancesystemforvehicletrackingand
classification," Intelligent Transportation
Systems, IEEE Transactions on, vol. 7, pp.
175-187, 2006.
21) H. Mao-Chi and Y. Shwu-Huey, "A real-time
and colorbased computer vision for traffic
monitoring system," in Multimedia and
Expo, 2004. ICME '04. 2004 IEEE
International Conference on, 2004, pp.
2119-2122 Vol.3.
22) Youssef ZINBI, YoussefCHAHIR,S.―Moving
object segmentation usingopticalflowwith
active contour model‖. IEEE Conferenceon
ICTTA, 2008,pp. 1-5.
23) . Ondr´uˇska and I. Posner, (2016)., ―Deep
tracking: Seeing beyond seeing using
recurrent neural networks,‖ in The
Thirtieth AAAI Conference on Artificial
Intelligence (AAAI), Phoenix, Arizona USA,
February 2016
24) P. Ondr´uˇska, J. Dequaire, D. Z. Wang, and I.
Posner, (2016)., ―End-to-end tracking and
semantic segmentation using recurrent
neural networks,‖ arXiv preprint
arXiv:1604.05091, 2016.
25) Jing Xin, Xing Du, Jian Zhang (2017), Deep
LearningForRobustOutdoorVehicleVisual
Tracking, Proceedings of the IEEE
International Conference on Multimedia
and Expo (ICME) 2017.*
26) W. Zhang, et al., "Moving vehicles detection
based on adaptivemotionhistogram,"Digit.
Signal Process., vol. 20, pp. 793-805, 2010.
27) W. Tao and Z. Zhigang, "Real time moving
vehicle detection and reconstruction for
improvingclassification,"inApplicationsof
Computer Vision (WACV), 2012 IEEE
Workshop on, 2012, pp. 497-502.
28) C. Yen-Lin, et al., "Real-time vision-based
multiple vehicle detection and tracking for
nighttime traffic surveillance," in Systems,
Man and Cybernetics, 2009. SMC 2009.
IEEE InternationalConferenceon,2009,pp.
3352-3358.
29) Witten, D. M. and Tibshirani, R. (2011).
Penalized Classification using Fisher’s
Linear Dis- criminant, Journal of the Royal
Statistical Society: Series B (Statistical
Methodology) 73(5): 753–772.
30) Sonka, M., Hlavac, V. and Boyle, R. (1999).
Image Processing, Analysis, and Machine
Vision, PWS Pub.
31) Han, F., Shan, Y., Cekander, R., Sawhney, H.
and Kumar, R. (2006). A TwoStage
Approach to People and Vehicle Detection
with Hog-Based SVM, Performance Metrics
for Intelligent Systems Workshop in
conjunction with the IEEE Safety, Security,
and Rescue Robotics Conference, pp.133–
140.
32) Ramakrishnan, V., Prabhavathy, A. K. and
Devishree, J. (2012). A Survey on Vehicle
DetectionTechniquesinAerialSurveillance,
International Journal of Computer
Applications 55(18).
33) Chen, Z., Pears, N., Freeman, M. and Austin,
J. (2009). Road vehicle classification using
support vector machines, Intelligent
Computing and Intelligent Systems, 2009.
ICIS 2009. IEEE International Conference
on, Vol. 4, IEEE, pp. 214–218.
34) Asha, G., Kumar, K. A. and Kumar, D. D. N. P.
(2012). A Real Time Video Object Tracking
Using SVM, International Journal of
Engineering Science and Innovative
Technology (IJESIT)
35) A.Albaqul ,H.G.Hamed ,A.Zaragoun Design
and Fabrication of a Smart Traffic Light

Control System Published 2012 Computer
Science International Conference on
Electronics, Computers and Artificial
Intelligence (ECAI), 2012.
36) Hazim Hamza, Prof. Paul Whelan, Night
TimeCarRecognitionUsingMATLAB,MEng
in Electronic Systems 2013 .
37) kzavya P Walad, Jyothi Shetty, Traffic Light
Control System Using Image Processing,
Vol.2, Special Issue 5, October 2014
38) Xiaoling Wang, Li-Min Meng, Biaobiao
Zhang, Junjie Lu, K,-L. Ju, A video-based
traffic violation detection system, 978-1-
4799-2565-0/13/$31.00 ©2013 IEEE.
39) Waing, Dr. Nyein Aye, On the Automatic
Detection System of Stop Line Violation for
Myanmar Vehicles (Car), Volume 1 -Issue4
November 2013
40) Md. Rifat Rayhan, Faysal, Mohammad , Md.
Taslim Reza, Improvement of a Traffic
System using Image and Video Processing,
Volume 2, Issue 3 (May. – Jun. 2013),

Vehicle Traffic Analysis using CNN Algorithm

More Related Content

Similar to Vehicle Traffic Analysis using CNN Algorithm (20)

More from IRJET Journal (20)

Recently uploaded (20)

Vehicle Traffic Analysis using CNN Algorithm