(1) Title of the project:

Dynamic Road Traffic Congestion and Density based Traffic Control Optimization Vehicle
Network and System Traffic Intelligent Communication using Image Processing Machine
Learning.

(2) Introduction: Giving purpose of research

Traffic congestion leads to a progressive reduction in traffic speeds, resulting in an increase in
journey times, fuel consumption, other operating costs, and environmental pollution contributing to
global warming. Traffic congestion has become a serious problem in recent days due to the
increased number of vehicles and the need for people to travel more for work and vacation. It may
bring stress to people, and makes them get stuck in traffic. It also makes people take longer to
reach their destination. These problems of congestion are suffered not only by the motorists but
also by the users of public transport, generally lower-income persons who have to pay higher fares
on account of congestion.

Another major consequence of traffic congestion is that an ambulance or a fire truck is unable to
respond in an awaited time. People may encounter trouble with their important work due to traffic
jams. This may lead to personal as well as professional drawbacks. Traffic congestion can also
cause an impact on the mind of a person because of the constant blowing of horns and frustration.
Various measures have has been proposed by the government to overcome the traffic jam problem
by imposing parking restrictions and adopting traffic signal rules. The existing traffic light control
approaches are time-dependent instead of the vehicle density on the road. It postulates manual
handling which causes congestion and longer waiting time. This hesitation in waiting time
produces traffic jams and creates more smoke emission. Thus, a traffic light control system based
on image processing is the optimal solution for traffic congestion.

The research work is designed to develop a density based dynamic traffic signal system. The signal
timing changes automatically on sensing the traffic density at the junction. Traffic congestion is a
severe problem in many major cities across the world and it has become a nightmare for the
commuters in these cities. Conventional traffic light system is based on fixed time concept allotted
to each side of the junction which cannot be varied as per varying traffic density. Junction timings

1
allotted are fixed. Sometimes higher traffic density at one side of the junction demands longer
green time as compared to standard allotted time.
The object detection in the traffic signal is processed and converted into frames then its threshold
is calculated based on which the contour has been drawn in order to calculate the number of
vehicles present in the area. After calculating the number of vehicles we will came to know in
which side the density is high based on which signals will be allotted for a particular side.

Due to the well-documented problems associated with inductive loops, most jurisdictions have
replaced many intersection loops with video image vehicle detection systems (VIVDS). While
VIVDS have overcome some of the problems with loops such as traffic disruption and pavement
degradation, they have not been as accurate as originally anticipated. The objective of this project
is to conduct evaluations of alternative detector technologies for application into the state’s traffic
signal systems. The research will include investigating the available detectors that could replace
loops or VIVDS through a literature search and agency contacts, followed by field and/or
laboratory investigations of promising technologies.

Deliverables will include a research report, a project summary report, and a detector selection
guide. Findings indicate that three detectors should be considered as alternatives to VIVDS for
signalized intersections – one is a radar detector and the other two are magnetic detectors. The
radar detector is only for dilemma zone detection and does not cover the stop line area. The other
two are point detectors, so their basic function would be for loop replacements. One is an intrusive
detector, requiring a short lane closure for installation and replacement. Field testing of
2
performance for all three detectors indicated they are worth considering as inductive loop or
VIVDS replacements.

Use of existing popular technologies for traffic congestion detection and management has been
explored by several researchers in recent times. However, most of the works either deal with
congestion detection strategies or address the issue of congestion control through signal
manipulation or other aids. To the best of my knowledge, no effort has been made so far to develop
a comprehensive system that automatically detects as well as controls traffic congestion based on
detected level of congestion. The aim of this work is to design a complete strategy for automatic
road traffic congestion detection in real-time and to devise an effective scheme for control/
management of congestion using video surveillance number of vehicles with the help of frames in
same area or lane.

In this project, discussed some of the most widely adopted technologies for congestion detection
and management, their limitations. Finally, we described the outline of our Active frames based
congestion detection and management scheme to overcome the existing problems. Further, we
have illustrated a object detection model using machine learning and image processing that we
have developed to generate different traffic congestion scenarios, to implement and test our
scheme for congestion detection and management.

Traffic Light Control:

Traffic lights are signals that are used to control the movement of vehicles on the road. This traffic
signal when effectively monitored can control the traffic on the road and avoid congestion. The
traffic light on the road comprises of 3 signals- red, yellow and green. People are made to hold
back for the green signal to further proceed. Delay in the red signal cause longer waiting time
because of congestion. CCTV cameras are installed at almost all junctions on the urban cities and
accident-prone and traffic-prone areas. This would enhance traffic maintenance and help police
book violators. With the help of these CCTV cameras, we can calculate the density of vehicles on
the road through faming and vehicle count. Image processing is the use of computer algorithms to
perform image processing on digital images. This technique controls the traffic light signals on the
road in order to avoid unnecessary traffic congestion.

3
(3) A brief review of the work already done in the field
Kaizhou Gao, Yicheng Zhang, Rong Su, Fajun Yang, Ponnuthurai Nagaratnam Suganthan, and
MengChu Zhou, [1] This paper addresses a traffic signal scheduling (TSS) problem in a
heterogeneous traffic network with signalized and non-signalized intersections. The objective is to
minimize the total network-wise delay time of all vehicles within a given finite-time window. First,
a novel model is proposed to describe a heterogeneous traffic network with signalized and non-
signalized intersections. Second, five meta-heuristics are implemented to solve the TSS problem.
Based on the problem characteristics, three local search operators and their ensemble are proposed.
Then, five meta-heuristics with such an ensemble are proposed to solve the TSS problem. Third,
experiments are carried out based on the real traffic data in the Jurong area of Singapore. The
performance of the ensemble of local search operators is verified. Ten algorithms, including five
meta-heuristics with and without the ensemble, are evaluated by solving 18 cases with different
scales. Finally, the algorithm with the best performance is compared against the currently used
traffic signal control strategies. The comparisons and discussions show the competitiveness of the
proposed model and meta-heuristics.

Palwasha W. Shaikh, Mohammed El-Abd, Mounib Khanafer, and Kaizhou Gao, [2] The rapid
development of urban cities coupled with the rise in population has led to an exponentially growing
number of vehicles on the roads for the latter to commute. This is adding to the already overbearing
problem of traffic congestion. Short term, costly and short-sighted solutions of road infrastructure
expansions are no longer suitable. One effective method of road resource allocation is focusing on
the widely used traffic signal controllers’ timing schedules. Searching for a suitable or an optimal
schedule for the prior via brute force to ease traffic congestion might not be the most elegant or
feasible solution. Nature-inspired algorithms including evolutionary and swarm intelligence
algorithms are gaining a lot of momentum. Many of these algorithms have been used in the last two
decades to address different applications in the smart city era including traffic signal control (TSC).
This paper conducts a comprehensive literature review on applications of evolutionary and swarm
intelligence algorithms to TSC. Surveyed work is categorized based on the set of decision
variables, optimization objective(s), problem modeling and solution encoding. The paper, based on
gaps identified by the conducted review, identifies promising future research directions and

4
discusses where the future research is headed.

Lisheng Jin, Mei Chen, Yuying Jiang, Haipeng Xia, [3] Traffic accidents are particularly serious at
a rainy day, night without street lamp, overcast, rainy night, foggy day and many other low
visibility conditions. Present vision driver assistance systems are designed to perform under good-
natured weather conditions. Classification is a methodology to identify the type of optical
characteristics for vision enhancement algorithms to make them more efficient. To improve
machine vision in bad weather situations, a multi-class weather classification method is presented
based on multiple weather features and supervised learning. Firstly, underlying visual features are
extracted from multi-traffic scene images, and then the feature was expressed as an eight-
dimension feature matrix. Secondly, five supervised learning algorithms are used to train
classifiers. The analysis shows that extracted features can accurately describe the image semantics
and the classifiers have high recognition accuracy rate and adaptive ability. The proposed method
provides the basis for further enhancing the detection of anterior vehicle detection during nighttime
illumination changes, as well as enhancing the driver's field of vision in a foggy day.

Jayani Withanawasam, Asoka Karunananda, [4] Transportation has been a crucial aspect of human
life for decades. Road traffic congestion is a critical socio-economic issue as it causes high fuel
consumption, waste of time, increased environmental pollution, and frustration and safety issues.
Major reason for increased road traffic congestion is the suboptimal use of available resources such
as time and road space due to the static traffic signal plan. Traffic density at multiple adjacent
intersections is not effectively considered in current traffic signal control plans. Novel perimeter
gating control mechanism based on the dynamic traffic conditions in multiple intersections is
proposed using multi-agent systems to improve the throughput of macroscopic traffic networks.
Multi-agent systems facilitate the communication and coordination between adjacent intersections.
Goal is to attain the emerging effect of minimizing the time loss due to traffic congestion over time
for a selected area. Simulation of Urban Mobility (SUMO) traffic simulation suite along with Java
Agent Development Environment (JADE) is used to assess the proposed perimeter gating control
mechanism. Experimental results show that the proposed method reduces the time loss by 34%
when compared with static traffic control method.

5
Xili Wan, Wentian Zhao, Xinjie Guan, Feng Ye and Guangwei Bai, [5] In urban area, fast growth
in the number of vehicles has led to a series of traffic problems, including traffic jams, high traffic
accident rates, etc. Efficient traffic signal control methods have been shown to be an essential way
to significantly mitigate traffic problems. In this poster, different from previous online methods,
authors propose a frame-based model and an efficient algorithm to solve the drawbacks of online
algorithm by scheduling the vehicles that have accumulated at the intersection over a period of
time. Preliminary experiments exhibit that the proposed algorithm could greatly improve the
throughput of the intersection.

Eduard Zadobrischi, Lucian-Mihai Cosovanu, and Mihai Dimian, [6] the studies carried out so far
try to highlight solutions that will streamline the traffic, as society revolves around transportation
and its symmetry. Current research highlights that the increased density of vehicles could be
remedied by dedicated short-range communications (DSRC) systems through communications of
the type vehicle-to-vehicle (V2V), vehicle-to-infrastructure (V2I) or vehicle-to-everything (V2X).
Authors can say that wireless communication technologies have the potential to significantly
change the efficiency and road safety, thus improving the efficiency of transport systems. An
important factor is to comply with the requirements imposed on the use of vehicle safety and
transport applications. Therefore, this paper focuses on several simulations on the basis of
symmetry models, implemented in practical cases in order to streamline vehicle density and reduce
traffic congestion. The scenarios aim at both the communication of the vehicles with each other and
their prioritization by the infrastructure, so authors can have a report on the efficiency of the
proposed models.

Jizhi Peng, Zhixiong Nan, Linhai Xu, Jingmin Xin, Nanning Zheng, [7] In this paper, inspired by
some recent works, authors propose a deep neural network model for joint object detection and
semantic segmentation. Given an image, an encoder-decoder convolution network extracts a set of
feature maps, these feature maps are shared by the detection branch and the segmentation branch to
jointly carry out the object detection and semantic segmentation. In the detection branch, we design
a PriorBox initialization mechanism to propose more object candidates. In the segmentation
branch, authors use the multi-scale atrous convolution to explore the global and local semantic
information in traffic scenes. Benefiting from the PriorBox Initialization Mechanism (PBIM) and

6
Multi-Scale Atrous Convolution (MSAC), our model presents the competitive performance. In the
experiments, authors widely compare with several recently-proposed methods on the public
Cityscapes dataset, achieving the highest accuracy.

Sakib B. Kibria, Mohammad S. Hasan, [8] One of the important practical applications of object
detection and image classification can be for security enhancement. If dangerous objects e.g.
knives can be identified automatically, then a lot of violence can be prevented. For this purpose,
various different algorithms and methods are out there that can be used. In this paper, four of them
have been investigated to find out which can identify knives from a dataset of images more
accurately. Among Bag of Words, HOG-SVM, CNN and pre-trained Alexnet CNN, the deep
learning CNN methods are found to give best results, though they consume significantly more
resources.

Manish Chaturvedi, Mandeep Kaur, Nitin Rakesh, Parma Nand, [9] The aim of the paper is to
develop a system that can recognize various targeted objects in real-time by incorporating some
computer vision techniques and advanced boosted Machine Learning algorithms to solve the
problem of object recognition with high accuracy. The task of object detection is a combination of
both, image classification and image localization. Image classification identifies the class of target
object whereas localization identifies its position inside the input image. Viola-Jones is an
algorithm that can work effectively for recognizing a large number of categories such as airplanes,
cars, faces, etc. It works on feature extraction and uses Haar-Like features to achieve the objective.
So, Viola-J ones is the algorithm that will be used for the development purpose of this paper. Also,
instead of using a normal image, the image segmentation technique will be used to create segments
in the image that will make it easy to analyze and process the image for further computations.

Shaif Choudhury, Soummyo Priyo Chattopadhyay, Tapan Kumar Hazra, [10] In this paper authors
would describe a vehicle detection technique that can be used for traffic surveillance systems.
Intelligent traffic surveillance systems equipped with electronic devices, work by communicating
with moving vehicles about traffic conditions, monitor rules and regulations and avoid collision
between cars. Therefore the first step in this process is the detection of cars. The system uses Haar
like features for vehicle detection, which is generally used for face detection. Haar feature-based

7
cascade classifiers are an effective object detection method first proposed by Viola and Jones. It's a
machine learning based technique which uses a set of positive and negative images for training
purpose. Results show this method is quite fast and effective in detecting cars in real time CCTV
footages.
Ahmed Fawzi Otoom, Hatice Gunes, Massimo Piccardi, [11] In this paper author address the
problem of abandoned object classification in video surveillance. Our aim is to determine (i) which
feature extraction technique proves more useful for accurate object classification in a video
surveillance context (scale invariant image transform (SIFT) key points vs. geometric primitive
features), and (ii) how the resulting features affect classification accuracy and false positive rates
for different classification schemes used. Objects are classified into four different categories: bag
(s), person (s), trolley (s), and group (s) of people. Our experimental results show that the highest
recognition accuracy and the lowest false alarm rate are achieved by building a classifier based on
our proposed set of statistics of geometric primitives' features. Moreover, classification
performance based on this set of features proves to be more invariant across different learning
algorithms.

Mudassir Khalil, Jianping Li, Abida Sharif, Jalaluddin Khan, [12] Traffic congestion in major cities
is very common problem. Every modern city wants to handle this problem. Being stuck in traffic
jam wastes so much time and money of citizens. Traffic congestion can be minimized by timely
inform about its happening to the concerned department and also to incoming traffic. For this
authors proposed an approach in which we use satellites to detect cities traffic congestion on real
time. This approach not only informs about congestion but also provides a broader view of cities
traffic. Plenty of satellites are available around earth. High resolution video is taken from a specific
point in space with satellites cameras. This video is further processed by using images processing
techniques, and congestion is detected in real time. In result section we proved this idea
successfully.

Shanming Lin, Jun Tang, Xuewu Zhang, Yanyun Lv, [13] Authors presents this paper algorithms
for vision-based detection, tracking and trajectory generation of vehicles in simple video sequences
of traffic scenes which are registered by a stationary camera. Processing is done by three steps:
improved vehicle detection method, faster tracking algorithm and new vehicle trajectory generation

8
technique. Vehicles are viewed as rectangular blocks with certain active action. The propounded
method is based on the correspondences between regions and vehicles, as the vehicles moving
through the video sequence. The vehicle trajectory generation technique, used for gathering
statistical traffic data for analysis in vehicle flow detection, is also described in detail.
Experimental results from highway scenes demonstrate the effectiveness of the vehicle detection
and tracking method.

Vaibhav Malviya, Rahul Kala, [14] Object & people tracking is an excellent field of computer
vision in which we have tried to detect and track multiple vehicles and people at outdoor traffic
environment and indoor office environment respectively. A background subtraction algorithm is
applied for vehicle detection. Kalman filter is used to predict the estimated position of every
vehicle in the next frame and updating of new track. Some vehicles are also detected in cluttered
scenes. Every moving vehicle is counted in video frames. Multiple face detection and tracking is an
also attractive field of computer vision. The faces are behaviourally very different to vehicles and
indoor scenarios are also very different to outdoor scenarios. Hence a different methodology to
track people is used. In this paper, authors have tried to detect and track face of multiple people on
two different datasets with different height of camera. Point feature is extracted and compared it in
the successive frames to track face of multiple persons. Every face is bounded by rectangular shape
with unique identity. Authors have also counted total number of faces in the frame sequences.

Hetal K. Chavda, Maulik Dhamecha, [15] Tracking people or moving objects across a PTZ camera
and maintaining a track within a camera is a challenging task in applications of video surveillance.
The goal of object tracking is segmenting a region of interest from a video scene and keeping track
of its motion, positioning and occlusion. The object detection and object classification are
preceding steps for tracking an object in sequence of images. Object detection is performed to
check existence of objects in video and to precisely locate that object. Then detected object can be
classified in various categories such as humans, vehicles, birds, floating clouds, swaying tree and
other moving objects. Object tracking is performed using monitoring objects' spatial and temporal
changes during a video sequence, including its presence, position, size, shape, etc. Object tracking
is used in several applications such as video surveillance, robot vision, traffic monitoring, Video in
painting and Animation.

9
Rakesh Chandra Joshi, Mayank Joshi, Adithya Gaurav Singh, Sanjay Mathur, [16] Video
surveillance is going through an enormous change in the last decade and lots of research is
undergoing in this area. As object tracking is an integral part of such systems, it becomes necessary
to review all state-of-the-art methods and approaches which belong to object detection,
classification, and tracking. This paper consists of all such methods in a classified manner. After
detecting object, classification is done in order to track that later. Object tracking is the method to
locate the target in subsequent frames of video. There are also some challenges in video
surveillance which make it complex. Object tracking is characterized on some parameters like
Illumination variation, co-ordinates matching, environmental issues, tracking object's variety; pose
variation, occlusion, and motion blur etc. A vast classification is considered for review and
complete analysis of all possible techniques so that each aspect is thus covered along with present
research undergoing in that area.

N. Kavitha, D. N. Chandrappa, [17] The compilation of high quality traffic data is one of the basic
concerns for real time traffic video surveillance. Since Video surveillance has figurative
applications such as security, traffic monitoring and law enforcement. Detecting and recognizing
the vehicle objects in a video is an imperative part of video surveillance systems, which is going to
guide in tracking of the detected objects and gathering decisive information. Vehicular detection
and classification is decisive as it can guide in traffic control and gathering of traffic info that can
be used in intelligent transportation systems. Vehicular classification poses a thought-provoking
problem as vehicles have high intra-class variation and comparatively low inter-class variation and
it is time consuming, varying performance under different weather conditions etc. Hence there is a
demand for employing innovative efficient data collection technique which will give the
advantageous continuous data collection with acceptable level of accuracy.

Angel Sánchez, Pedro D. Suárez, Aura Conci, Éldman Nunes, [18] Imaging hardware and video
processing techniques offer advantages for traffic monitoring and surveillance. As these
experiments show, using appropriate road and vehicle modeling and strong vehicle detection and
tracking algorithms offers a good trade-off between correct dynamic vehicle identification and a
real-time frame rate. Further, heuristic rules can help analyze and solve difficult traffic situations.

10
Akshay Mangawati, Mohana, Mohammed Leesan, H. V. Ravish Aradhya, [19] Object tracking is
one of the most critical areas of research due to change in motion of object. Specifically, feature
selection places a vital role in object tracking. It has wide range of applications such as object
detection, traffic control, human-computer interaction, gesture recognition, video surveillance.
Many approaches focus on the tracking algorithm to smoothen the video sequence in order to
overcome the issues of tracking, related to object movement and appearance. These methods uses
object shape, colour, texture, object of interest and motion in multi direction in tracking for video
surveillance applications. This paper elaborate the exhaustive survey of various object tracking
algorithms under different environmental conditions and identified efficient algorithm in various
types of tracking. In this paper objects are tracked based on colour, motion of single and multiple
objects (vehicles) are detected and counted in multiple frames. Further single algorithm may be
designed for object tracking by considering shape, colour, texture, object of interest, motion of
object in multi direction.

Garima Mathur, Devendra Somwanshi, Mahesh M. Bundele, [20] The work proposed in this paper
was defined to meet following objectives: to create a video sequence for moving object tracking; to
design an experimental set up for moving object detection; and, to design and implement moving
object tracking algorithm, The proposed algorithm was implemented on a captured video sequence.
Object was detected first according to the image input, and then tracked in subsequent frames. The
experimental implementation could perform the object tracking without missing any frame and
could successfully overlay bounding box. It could successfully generate an image sequence after
the complete implementation of Mean Shift Algorithm. The performance of algorithm was checked
by successfully tracking the user defined object and performing the overlay function in the
identified object.

Chinmaya Kumar Sahu, Sushree Barsa Pattnayak, Susantini Behera, Manas Ranjan Mohanty, [21]
Automatic number plate detection and analysis is a general monitoring strategy used by a large
number of city vehicles to enhance traffic management, routing, traffic control, toll collection, and
regulation and protection of highway law. ANPR approach can be applied according to different
methodologies. This job can be scanned, executed and compared. This proposed work is carried out

11
in real-time application using YOLO v3 for the identification and recognition of plate numbers. In
this study, a comparative method for ANPR has been demonstrated. Traditional approaches were
focused on contouring, segmentation, edge detection processes which gave less accuracy but here
tried to implement YOLO v3 technique that will give more accurate results for Indian license plate
detection in real-time.

Ling Hu, Qiang Ni, [22] Automated object detection algorithm is an important research challenge
in intelligent urban surveillance systems for Internet of Things (IoT) and smart cities applications.
In particular, smart vehicle license plate recognition and vehicle detection are recognized as core
research issues of these IoT-driven intelligent urban surveillance systems. There are key techniques
in most of the traffic related IoT applications, such as road traffic real-time monitoring, security
control of restricted areas, automatic parking access control, searching stolen vehicles, etc. In this
paper, authors propose a novel unified method of automated object detection for urban surveillance
systems. Authors use this novel method to determine and pick out the highest energy frequency
areas of the images from the digital camera imaging sensors, that is, either to pick the vehicle
license plates or the vehicles out from the images. Our proposed method can not only help to detect
object vehicles rapidly and accurately, but also can be used to reduce big data volume needed to be
stored in urban surveillance systems.

Huansheng Song, Haoxiang Liang, Huaiyu Li, Zhe Dai & Xu Yun, [23] Intelligent vehicle

detection and counting are becoming increasingly important in the field of highway management.
However, due to the different sizes of vehicles, their detection remains a challenge that directly
affects the accuracy of vehicle counts. To address this issue, this paper proposes vision-based
vehicle detection and counting system. A new high definition highway vehicle dataset with a total
of 57,290 annotated instances in 11,129 images is published in this study. Compared with the
existing public datasets, the proposed dataset contains annotated tiny objects in the image, which
provides the complete data foundation for vehicle detection based on deep learning. In the
proposed vehicle detection and counting system, the highway road surface in the image is first
extracted and divided into a remote area and a proximal area by a newly proposed segmentation
method; the method is crucial for improving vehicle detection. Then, the above two areas are
placed into the YOLOv3 network to detect the type and location of the vehicle. Finally, the vehicle

12
trajectories are obtained by the ORB algorithm, which can be used to judge the driving direction of
the vehicle and obtain the number of different vehicles. Several highway surveillance videos based
on different scenes are used to verify the proposed methods.

Myungeun Eom & Byung-In Kim, [24] The intersection traffic signal control problem (ITSCP) has
become even more important as traffic congestion has been more intractable. The ITSCP seeks an
efficient schedule for traffic signal settings at intersections with the goal of maximizing traffic flow
while considering various factors such as real-time strategies; signal timing constraints, rapid
developments in traffic systems, and practical implementation. Since the factors constituting the
ITSCP exhibit stochastically complicated interactions, it is essential to identify these factors to
propose solution methods that can address this complexity and still be practically implemented.

Ach. Maulana Habibi Yusuf, Rahadian Yusuf, [25] Traffic light regulation becomes important as
the number of vehicles increases in densely populated areas. Adaptive traffic light allows traffic
lights to adjust the flow of vehicles according to the density of the vehicle on each section. This
paper proposed the use of agent-based modeling system to control traffic light and focus on
modeling based on unity3D platform using fuzzy logic control. The test was carried out on three
conditions, namely crowded, moderate, and smooth conditions. The system performance
parameters are the average waiting time, the average of vehicles stopping, and the average speed of
the vehicles. Simulation results show that adaptive traffic light can reduce waiting time, reduce the
number of vehicles stopping at an intersection, and increase vehicles speed compared to fixed time
traffic light.

(4) Noteworthy contributions in the field of proposed work

The main contribution of the proposed framework is as follows:
1) A framework architecture is proposed for research work that includes frame/segment level
feature extraction, dynamic feature generation, feature dimension reduction, and regression;
2) Conducting a comprehensive literature review on optimizing the Traffic Signal Control
problem using EA and SI approaches.
3) Providing insights into future research directions based on identified gaps in the current

13
 research.
4) Different traffic networks located in highly congested areas should be considered in order to
monitor the performance and efficiency of the proposed strategy.
5) To improve the performance and accuracy of proposed algorithms.
6) Traffic signal setting and planning in a heterogeneous traffic network should be researched.

(5) Proposed methodology during the tenure of the research work

Traffic congestion at junctions is a serious issue on a daily basis. The prevailing traffic light
controllers are unable to manage the different traffic flows. Most of the current systems operate
on a timing mechanism that changes the signal after a particular interval of time. This may
cause frustration and result in motorist's time waste. Traffic congestion is a major problem in
the currently existing systems. Delays, safety, parking, and environmental problems are the
main issues of current traffic systems that emit smoke and contribute to increasing Global
Warming. Sensor-based systems reduce the waiting time and maximize the total number of
vehicles that can cross an intersection. Our proposed system can control the traffic lights based
on image processing without the need for traffic police. This can reduce congestion, delay,
road accidents, need for manpower. Under image processing, we use sub techniques like RGB
to Gray conversion, Image resizing, Image Enhancement, Edge detection, Image matching, and
Timing allocation. A real-time image is captured for every 1 second. After edge detection
procedure for both reference and real-time images, these images are compared using CCTV
surveillance. Then the amount of traffic is detected and the details are stored in the server. Four
cameras are placed on respective roads and these cameras are used to capture images to
analyze traffic density. Then the traffic signals are decided according to the density of traffic.
Our technique can be effective to combat traffic on Indian Roads. A lot of time can be saved
by deploying this system and also it conserves a lot of resources as well as the economy.

A. Existing Methods:
There are numerous procedures proposed already for traffic monitoring to avoid the traffic jam
on the road. The previously practiced traffic control system in most of the cities is based on
fixed time control or manual control. However, the new implementations in a few cities focus
on the implementation of the 'Adaptive Traffic Control System' (ATCS). In this system, with

14
 the help of traffic detectors, the traffic density is obtained from all approaches of 1 intersection
of the road. The traffic light control time-dependent system has a fixed period to switch traffic
between different directions. The traffic lights turn ON and OFF by itself based on the timer
value. Consequently, the vehicles are made to wait for a longer period even on the empty road,
or if the traffic density is very less. Using sensors is another way to control traffic by detecting
the number of vehicles. Another most common strategy to control traffic is manual controlling.
Manual controlling involves the need for traffic police. The traffic signals can be changed by
the police manually based on the vehicles and traffic flow on the road by himself.

B. Issues in Existing Methods:

A time-based traffic control system is one of the common methodologies or techniques to

control traffic. But this methodology is contingent on time rather than on density. On account
of this, the vehicles have to encounter an erratic delay in waiting time. The vehicles are made
to wait even on the empty road in a time-based traffic control system which can again
contribute congestion or traffic jam. Sensors can be used to detect vehicles and control traffic
accordingly. But even in this method, the time is wasted by the green light on an empty road.
Manual controlling engages the need for traffic police to change the traffic signals accordingly.
This method also requires manpower. Although abundant methods are present already to
control traffic, these methods can still bring on congestion due to disparate reasons. A traffic
light control system based on the image processing technique can handle the problem of traffic
congestion more effectively.

C. Proposed Methodology:

The proposed system adopts the traffic light signal according to the real-time traffic density
using image processing techniques. This model uses CCTV cameras to sense the changing
traffic patterns around the traffic signal and manipulates the traffic signals accordingly by
triggering the signals to the traffic control system. The modifications in traffic jams directly
rely on the control of traffic flow, and hence on the traffic signal. In view of the fact that
CCTV cameras are operational at almost every junction these days, it can be used to control

15
 traffic and keep clear of unwanted congestion.

Fig.1: Proposed System Model

The system that we put forward consists of a CCTV camera that is used to capture the
reference image and the actual image. These images are stored in a database. It is later
retrieved from database and Image process techniques are applied and are feed into Adaptive
Traffic Control System (ATCS).

D. Functional Architecture:

In our proposed system, the reference image of an empty road is taken and the real-time traffic
on a road is captured as an actual image. They are stored in the database and the data is
extracted using data extraction tools. Image processing techniques like RGB to Grayscale
conversion, Image resizing, Image enhancement, Edge detection and Framing of objects are
used and applied to both the images. Image matching for both the images is done and the
percentage of matching is done using the machine learning algorithm.

16
 Fig.2: Functional Architecture

List of Modules:

1. Capturing and storing in DB

2. Extraction of images

3. Image processing techniques

4. Object Detection and framing

5. Counting the number of frames as per lane

6. Navigation and control of signals using time allocation

E. Deep Convolutional Neural Network:

CNN is one of the main categories to do image recognition, image classification. Object
detection, face recognition, emotion recognition etc., are some of the areas where CNN are
widely used. CNN image classification takes an input image, process it and classify it under

17
 certain categories. CNN is a neural network that has one or more convolutional layers.

 Step 1: Dataset containing object images along with reference frames is fed into the
System.

 Step 2: Now import the required libraries and build the model.

 Step 3: The convolutional neural network is used which extracts image features f pixel by
pixel.

 Step 4: Matrix factorization is performed on the extracted pixels. The matrix is of m x n.

 Step 5: Max pooling is performed on this matrix where maximum value is selected and
again fifixed into matrix.

 Step 6: Normalization is performed where the every negative value is converted to zero.

 Step 7: To convert values to zero rectifified linear units are used where each value is
ffiltered and negative value is set to zero.

 Step 8: The hidden layers take the input values from the visible layers and assign the
weights after calculating maximum probability.

(6) Expected outcome of the proposed work (Objectives)

A. Objectives:
The objective of this work is to design proposed system which has:
• To solve the traffic congestion problem, an designed a framework for a dynamic and
automatic traffic light control system and developed a software based model with codes in to
help build the system on hardware.
• To implement a system in that first signal timing changes automatically on sensing the traffic
density with the help of CCTV footage object detection and counting the frames.
• To design an system, in case of any emergency vehicle like ambulance, fire brigade etc.,
requiring priority are built in with control to override the set timing by providing
instantaneous green signal in the desired direction while blocking the other lanes by red signal
for some time.
• To propose dynamic time-based coordination schemes where the green signal time of the

18
 traffic lights is assigned based on the present conditions of traffic.

B. Outcome of the proposed work:

The outcomes of research work are as follows:
• The aim of the project is to solve traffic congestion which is a severe problem in many
modern cities all over the world.
• During normal time the signal timing changes automatically on sensing the traffic density at
the junction but in case of any emergency vehicle like ambulance, fire brigade etc., requiring
priority are built in with system control to override the set timing by providing instantaneous
green signal in the desired direction while blocking the other lanes by red signal for some
time.
• It reduces the possibilities of traffic jams caused by traffic lights to an extent.

C. Scope of the proposed work:

Consequently, the scope of this research work can be summarized as:
• By using this system configuration we can reduce the possibilities of traffic jams caused by
traffic lights to an extent.
• The number of passing vehicles on the road decides the density range of traffics and on the
basis of vehicle count system decides the traffic light delays.

19
(7) References

[1] Kaizhou Gao, Yicheng Zhang, Rong Su, Fajun Yang, Ponnuthurai Nagaratnam Suganthan,
and MengChu Zhou, “Solving Traffic Signal Scheduling Problems in Heterogeneous Traffic
Network by Using Meta-Heuristics”. IEEE Transactions on Intelligent Transportation
Systems, 2018.
[2] Palwasha W. Shaikh, Mohammed El-Abd, Mounib Khanafer, and Kaizhou Gao, “A Review
on Swarm Intelligence and Evolutionary Algorithms for Solving the Traffic Signal Control
Problem”. IEEE Transactions on Intelligent Transportation Systems, 2020
[3] Lisheng Jin, Mei Chen, Yuying Jiang, Haipeng Xia, “Multi-Traffic Scene Perception Based
on Supervised Learning”. IEEE Access ( Volume: 6), January 2018
[4] Jayani Withanawasam, Asoka Karunananda, “Multi-agent based road traffic control
optimization”. 2017 IEEE 20th International Conference on Intelligent transportation
Systems (ITSC), March 2018
[5] Xili Wan, Wentian Zhao, Xinjie Guan, Feng Ye and Guangwei Bai, “Performance
Guaranteed Traffic Signal Control with Frame-based Algorithm”. 2018 15th Annual IEEE
International Conference on Sensing, Communication, and Networking (SECON)
[6] Eduard Zadobrischi, Lucian-Mihai Cosovanu, and Mihai Dimian, “Traffic Flow Density
Model and Dynamic Traffic Congestion Model Simulation Based on Practice Case with
Vehicle Network and System Traffic Intelligent Communication”. This article belongs to the
Special Issue Symmetry and IoT Intelligence in the Post Pandemic Economy, July 2020
[7] Jizhi Peng, Zhixiong Nan, Linhai Xu, Jingmin Xin, Nanning Zheng, “A Deep Model for Joint
Object Detection and Semantic Segmentation in Traffic Scenes”. 2020 International Joint
Conference on Neural Networks (IJCNN)
[8] Sakib B. Kibria, Mohammad S. Hasan, “An analysis of Feature extraction and Classification
Algorithms for Dangerous Object Detection”. 2017 2nd International Conference on
Electrical & Electronic Engineering (ICEEE)
[9] Manish Chaturvedi, Mandeep Kaur, Nitin Rakesh, Parma Nand, “Object Recognition Using
Image Segmentation”. 2020 Sixth International Conference on Parallel, Distributed and Grid
Computing (PDGC)

20
[10] Shaif Choudhury, Soummyo Priyo Chattopadhyay, Tapan Kumar Hazra, “Vehicle Detection
and Counting using Haar Feature-Based Classifier”. 2017 8th Annual Industrial Automation
and Electromechanical Engineering Conference (IEMECON)
[11] Ahmed Fawzi Otoom, Hatice Gunes, Massimo Piccardi, “Feature extraction techniques for
abandoned object classification in video surveillance”. 2008 15th IEEE International
Conference on Image Processing.
[12] Mudassir Khalil, Jianping Li, Abida Sharif, Jalaluddin Khan, “Traffic Congestion Detection
by Use of Satellites View”. 2017 14th International Computer Conference on Wavelet Active
Media Technology and Information Processing (ICCWAMTIP)
[13] Shanming Lin, Jun Tang, Xuewu Zhang, Yanyun Lv, “Research on Traffic Moving Object
Detection, Tracking and Track-Generating”. 2009 IEEE International Conference on
Automation and Logistics.
[14] Vaibhav Malviya, Rahul Kala, “Tracking Vehicle and Faces: Towards Socialistic Assessment
of Human Behaviour”. 2018 Conference on Information and Communication Technology
(CICT).
[15] Hetal K. Chavda, Maulik Dhamecha, “Moving Object Tracking using PTZ Camera in Video
Surveillance System”. 2017 International Conference on Energy, Communication, Data
Analytics and Soft Computing (ICECDS)
[16] Rakesh Chandra Joshi, Mayank Joshi, Adithya Gaurav Singh, Sanjay Mathur, “Object
Detection, Classification and Tracking Methods for Video Surveillance: A Review”. 2018 4th
International Conference on Computing Communication and Automation (ICCCA)
[17] N. Kavitha, D. N. Chandrappa, “Comparative Review on Video Based Vehicular Traffic Data
Collection for Intelligent Transport System”. 2017 International Conference on Recent
Advances in Electronics and Communication Technology (ICRAECT)
[18] Angel Sánchez, Pedro D. Suárez, Aura Conci, Éldman Nunes, “Video-Based Distance Traffic
Analysis: Application to Vehicle Tracking and Counting”. Computing in Science &
Engineering (Volume: 13, Issue: 3, May-June 2011)
[19] Akshay Mangawati, Mohana, Mohammed Leesan, H. V. Ravish Aradhya, “Object Tracking
Algorithms for Video Surveillance Applications”. 2018 International Conference on
Communication and Signal Processing (ICCSP)

21
[20] Garima Mathur, Devendra Somwanshi, Mahesh M. Bundele, “Intelligent Video Surveillance
based on Object Tracking”. 2018 3rd International Conference and Workshops on Recent
Advances and Innovations in Engineering (ICRAIE)
[21] Chinmaya Kumar Sahu, Sushree Barsa Pattnayak, Susantini Behera, Manas Ranjan Mohanty,
“A Comparative Analysis of Deep Learning Approach for Automatic Number Plate
Recognition”. 2020 Fourth International Conference on I-SMAC (IoT in Social, Mobile,
Analytics and Cloud) (I-SMAC).
[22] Ling Hu, Qiang Ni, “IoT-Driven Automated Object Detection Algorithm for Urban
Surveillance Systems in Smart Cities”. IEEE Internet of Things Journal (Volume: 5, Issue: 2,
April 2018)
[23] Huansheng Song, Haoxiang Liang, Huaiyu Li, Zhe Dai & Xu Yun, “Vision-Based Vehicle
Detection and Counting System using Deep Learning in Highway Scenes”. European
Transport Research Review volume 11, Article number: 51 (2019)
[24] Myungeun Eom & Byung-In Kim, “The Traffic Signal Control Problem for Intersections: A
Review”. European Transport Research Review volume 12, Article number: 50 (2020)
[25] Ach. Maulana Habibi Yusuf, Rahadian Yusuf, “Adaptive Traffic Light Controller Simulation
for Traffic Management”. 2020 6th International Conference on Interactive Digital Media
(ICIDM)

(8) List of published papers of the candidate

1. Depression Scale Recognition over Fusion of Visual and Vocal Expression using
Artificial Intellectual Method.
International Journal of Recent Technology and Engineering (IJRTE) is online, open access,
peer reviewed, periodical international journal. (IJRTE-Scopus Indexed)
– First Author: Pratibha Gayke*, Department of Computer Science & Engineering, Dr.
APJ Abdul Kalam University, Indore, India. Email: [email protected]
– Second Author: Rohini S. Kulkarni, Department of Computer Science and
22
 Engineering, KLE Institute of Technology, Hubli, Karnataka, India. Email:
[email protected]

2. Transforming Healthcare Industry By Implementing Cloud Computing.

International Journal of Trend in Scientific Research and Development - ISSN: 2456-6470.
IN. IJTSRD is an International Peer Reviewed, Open Access Journal.
– First Author: Rohini S. Kulkarni*, Department of Computer Science and
Engineering, KLE Institute of Technology, Hubli, Karnataka, India. Email:
[email protected]
– Second Author: Pratibha Gayke, Department of Computer Science & Engineering,
Dr. APJ Abdul Kalam University, Indore, India. Email: [email protected]

23