Page 1 of 14 - Cover Page Submission ID trn:oid:::9380:83833245

DEEP LEARNING WITH IMAGE PROCESSING FOR OBJECT

DETECTION IN TRAFFIC CAMERAS.pdf
Hindustan Institute of Technology and Science, Kancheepuram

Document Details

Submission ID

trn:oid:::9380:83833245 10 Pages

Submission Date 3,924 Words

Feb 28, 2025, 11:34 AM GMT+5:30

22,917 Characters

Download Date

Feb 28, 2025, 11:36 AM GMT+5:30

File Name

DEEP LEARNING WITH IMAGE PROCESSING FOR OBJECT DETECTION IN TRAFFIC CAMERAS.pdf

File Size

4.3 MB

Page 1 of 14 - Cover Page Submission ID trn:oid:::9380:83833245

 Page 2 of 14 - Integrity Overview Submission ID trn:oid:::9380:83833245

28% Overall Similarity

The combined total of all matches, including overlapping sources, for each database.

Filtered from the Report

Bibliography

Quoted Text

Small Matches (less than 14 words)

Match Groups Top Sources

13 Not Cited or Quoted 28% 27% Internet sources

Matches with neither in-text citation nor quotation marks
17% Publications
0 Missing Quotations 0% 7% Submitted works (Student Papers)
Matches that are still very similar to source material

0 Missing Citation 0%
Matches that have quotation marks, but no in-text citation

0 Cited and Quoted 0%

Matches with in-text citation present, but no quotation marks

Integrity Flags
0 Integrity Flags for Review
Our system's algorithms look deeply at a document for any inconsistencies that
No suspicious text manipulations found. would set it apart from a normal submission. If we notice something strange, we flag
it for you to review.

A Flag is not necessarily an indicator of a problem. However, we'd recommend you

focus your attention there for further review.

Page 2 of 14 - Integrity Overview Submission ID trn:oid:::9380:83833245

 Page 3 of 14 - Integrity Overview Submission ID trn:oid:::9380:83833245

Match Groups Top Sources

13 Not Cited or Quoted 28% 27% Internet sources

Matches with neither in-text citation nor quotation marks
17% Publications
0 Missing Quotations 0% 7% Submitted works (Student Papers)
Matches that are still very similar to source material

0 Missing Citation 0%
Matches that have quotation marks, but no in-text citation

0 Cited and Quoted 0%

Matches with in-text citation present, but no quotation marks

Top Sources
The sources with the highest number of matches within the submission. Overlapping sources will not be displayed.

1 Internet

link.springer.com 5%

2 Internet

assets.researchsquare.com 5%

3 Internet

www.researchgate.net 4%

4 Internet

web.archive.org 3%

5 Internet

pubmed.ncbi.nlm.nih.gov 3%

6 Internet

etj.uotechnology.edu.iq 2%

7 Internet

www.ijraset.com 2%

8 Internet

www.mathworks.com <1%

9 Submitted works

Coventry University on 2023-04-10 <1%

10 Internet

turcomat.org <1%

Page 3 of 14 - Integrity Overview Submission ID trn:oid:::9380:83833245

 Page 4 of 14 - Integrity Overview Submission ID trn:oid:::9380:83833245

11 Internet

fastercapital.com <1%

12 Publication

H.C. Nguyen, H. Xu, M. Nazem, R. Sousa, J. Kowalski, Q. Zhao. "Deep-learning surr… <1%

13 Submitted works

uaveiro on 2024-10-28 <1%

Page 4 of 14 - Integrity Overview Submission ID trn:oid:::9380:83833245

 Page 5 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

DEEP LEARNING WITH IMAGE PROCESSING FOR OBJECT

DETECTION IN TRAFFIC CAMERAS

Santhosh D, Mr.Prabu
MCA Computer Applications, Assistant Professor(SS),
Hindustan Institution of Technology and Department of Computer Applications,
Science, Hindustan Institution of Technology and
Science,
Chennai, Tamilnadu, India.
Chennai, Tamilnadu
[email protected]

ABSTRACT accurate vehicle counting, which can be achieved

The rapid advancements in machine through object detection techniques. This paper
learning (ML) and image processing have proposes an approach for vehicle counting in
13 traffic scenes using Convolutional Neural
revolutionized traffic monitoring systems,
particularly in the area of object detection using Networks (CNNs), a class of deep learning
traffic cameras. This paper explores the algorithms well-suited for image recognition
integration of machine learning algorithms with tasks system is designed to detect and count
image processing techniques to develop efficient vehicles in real-time from traffic camera footage.
and accurate object detection systems for real- The results demonstrate the effectiveness of
time traffic surveillance Image processing CNN-based models for vehicle counting in
techniques, including image filtering, edge challenging conditions, achieving high accuracy
detection, and feature extraction, enhance the and fast inference speeds suitable for deployment
quality of input data, enabling better detection in real-time traffic monitoring systems.
Keywords: Object Detection,Traffic
accuracy. These object detection systems offer a
Cameras,Machine Learning (ML),Image
wide range of applications, such as traffic flow
Processing, Computer Vision,Deep Learning,YOLO
monitoring, vehicle counting, accident detection,
(You Only Look Once),Vehicle Detection
traffic violation enforcement, and pedestrian
I.INTRODUCTION
safety. Additionally, they play a crucial role in
With the increasing demand for intelligent
the development of smart transportation systems
transportation systems (ITS), accurate
(ITS) that can optimize traffic management and
monitoring of road traffic has become a crucial
improve road safety. The combination of
task. The ability to detect and count vehicles in
machine learning with image processing offers a
real-time is essential for various applications,
scalable and robust solution for real-time object
including traffic flow analysis, congestion
detection, contributing to more efficient, safer,
management, urban planning, and autonomous
and smarter traffic management systems. Object
vehicle navigation. Traditionally, vehicle
detection plays a crucial role in a wide range of
counting was performed manually or using
applications, including intelligent transportation
simpler computer vision techniques. The
systems (ITS), autonomous vehicles, and traffic
complexity and scale of traffic data have
monitoring. One of the key tasks in ITS is

Page 5 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 6 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

increased, more sophisticated approaches are commonly used to pre-process the raw image
leveraging deep learning, particularly data from traffic cameras, making it suitable for
Convolutional Neural Networks (CNNs), have further analysis by machine learning models.
become increasingly important. Convolutional
Neural Networks (CNNs), a type of deep II. LITERATURE REVIEW
9 learning model, have proven to be highly
In this paper [1] Traffic monitoring and
effective for image classification and object
management are crucial for improving road
detection tasks due to their ability to
safety and optimizing urban mobility. Traditional
automatically learn hierarchical features from
surveillance systems rely on manual observation,
raw image data.
which is time-consuming and inefficient. This
paper explores the application of machine
Detection of multiple moving vehicles in a video
learning techniques for object detection and
sequence. Detecting the helmet for safe and
classification in traffic camera footage. The
10 secure travelling Tracking of the detected
11 system utilizes deep learning models such as
vehicles. Identification of Vehicle types.
Convolutional Neural Networks (CNNs), YOLO
Counting the total number of vehicles passing in
(You Only Look Once), and Faster R-CNN to
videos.
detect and classify objects, including vehicles,
pedestrians, and traffic signs, in real-time.
Detection of multiple moving vehicles in a video
sequence. The rise of intelligent transportation
4 In this paper[2] In this challenging evolution, the
systems (ITS) has brought a significant
primary task in detecting the objects requires a
transformation in how traffic monitoring and
computer vision that deals over indoor and
management are approached. One key area
outdoor classes. Over the past decade, this zeal
where technology plays a pivotal role is in the
require more attention. Previous implementation
detection and analysis of objects on roads,
techniques include in object recognition with a
especially through the use of traffic cameras.
strategy of single labelling. Aim and Objectives:
These systems, coupled with machine learning
In this regard, a multi-label method using
(ML) and image processing techniques, have
machine learning and image technologies, and
made it possible to automatically detect vehicles,
accurate reply can be acknowledged built on its
pedestrians, cyclists, and other relevant objects
accuracy and efficiency. In the proposed work,
in real-time, improving traffic flow, safety, and
12
we solve the existing system problem by using
overall management. Machine learning (ML) is
classification/clustering techniques that are used
a subgroup of artificial intelligence (AI) that
to reduce the recognize time of multi objects in
enable system to automatically learn from data
less time with best time complexities. Model:
and make prediction or decisions without being
The model used to assist the visually impaired
explicitly programmed. In the context of object
people can independently recognize objects
detection, ML algorithms can be trained on large
which are near to them.
datasets of labeled images to recognize and
classify various objects present in traffic scenes.
1 In this paper [3] Goals. The present study was
Image processing, on the other hand, involves designed to build a prototyping and develop
the manipulation and analysis of visual data algorithms that allow the detection,
(images or video). Techniques such as edge classification, and movement of objects of a
detection, image filtering, and feature extraction robotic arm of 4 DOF with the following

Page 6 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 7 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

1 technologies: ArmUno arm structure, Raspberry anomalies include standing and traveling in
Pi 3 B+, PiCam 2.1, driver PCA9685 for reverse direction. Images are captured using
servomotors, Opencv3, and python. Another goal CCTV cameras from front and rear side of the
was to measure the effectiveness of prediction vehicle. This capability makes the results robust
and classification of objects photographed by the to the variations in operational and
robotic arm, using machine learning with the environmental conditions. Multiple consecutive
KNN classifier method.Methodology. The frames are acquired for motion detection.
generation of a dataset of 800 photographic Features such as edges and license plate corner
images was proposed, in 4 categories: volumetric locations are extracted for tracking purposes.
geometric shapes conformed by 200 images each Direction of the traffic flow is obtained from the
one of them. With this, processing techniques trained classifier. K-nearest neighbor is chosen
were applied to the image captured by the camera as the classifier model. The proposed method is
to detect the object in the image: Grayscale evaluated on a public highway and promising
filtering, Gaussian filtering, and threshold.Then, detection results are achieved.
2 the characteristics of the object were obtained In this paper[6] The usage of machine learning
through the first two invariant moments of HU, and deep learning algorithms have necessitated
and finally, the machine learning method KNN Artificial Intelligence'. AI is aimed at automating
was applied to predict, that the image captured things by limiting human interference. It is
by the robotic arm belongs or not to a certain widely used in IT, healthcare, finance, and
category. agriculture. It is achieved through several deep
learning algorithms that reflect the human brain's
In this paper[4] Object Detection Using Machine intelligence. These AI algorithms can be
Learning Common machine learning techniques manipulated according to changing needs and
8 include: Aggregate channel features (ACFs) improved efficiency. This paper tries to utilize
Support vector machine (SVM) classification the developments made in AI technology to
using histograms of oriented gradient (HOG) classify the images and recognize the objects
features. The Viola-Jones algorithm for human present in them. One widely used AI algorithm is
7 face or upper body detection. The detection of CNN (Convolutional Neural Networks). The
moving objects in videos and video surveillance, CNN is a deep learning-based algorithm that
which are both significant and difficult tasks in consists of various layers that extract and filters
many computer vision applications, are covered the parameters present in the images. Some
in this chapter\'s review and systematic additional layers of ResNet50 and the CNN
investigation. Detection algorithms for humans, algorithm are used to extract the parameters to
cars, threats, and security, for example. One of improve image recognition accuracy. The image
the most difficult research topics in computer dataset taken for training and testing the
vision right now is video surveillance in dynamic proposed model is imageNet. The images are
environments, especially for people, vehicles, initially processed before sending them to the
and specific objects in cases of security. proposed model.

5
3 In this paper[5] Anomaly detection is an In this paper[7] Road condition monitoring is
important part of an Intelligent Transportation essential for improving traffic safety and
System. In this study, image processing and reducing accidents. Machine learning methods
machine learning techniques are used to detect have recently gained prominence in the
anomalies in vehicle movements. These practically important task of controlling road

Page 7 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 8 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

5 surface quality. Several systems have been visual senses is astonishing. However, robots
proposed using sensors, especially have many substantial obstacles in detecting and
accelerometers present in smartphones due to identifying objects. Neural networks are a
their availability and low cost. However, these suggested solution from the field of computer
methods require practitioners to specify an exact science to address this problem. Additionally, it
set of features from all the sensors to provide is sometimes known as “Artificial Neural
more accurate results, including the time, Networks.“ This research focuses mostly on
frequency, and wavelet-domain signal features. It dynamic things, which are objects in motion. The
is important to know the effect of these features system is meant to perform object detection and
change on machine learning model performance static object identification to achieve its specified
in handling road anomalies classification tasks. purposes.
III. METHODOLOGY
6 In this paper[8] Moving objects detection, type
recognition, and traffic analysis in videobased 3.1 EXISTING SYSTEM
surveillance systems is an active area of research The existing system for object detection
which has many applications in road traffic in traffic cameras primarily relies on traditional
monitoring. This paper is on using classical computer vision techniques and sensor-based
approaches of image processing to develop an methods. These systems use motion detection
efficient algorithm for computer vision based on algorithms, background subtraction, and edge
traffic surveillance system that can detect and detection to identify moving objects in traffic
classify moving vehicles, besides serving some scenes. Additionally, some implementations
other traffic analysis issues like finding vehicles depend on hardware-based solutions like
speed and heading, tracking specified vehicles, inductive loop sensors, LiDAR, or radar for
and finding traffic load. vehicle detection. However, these approaches
often suffer from limitations such as low
In this paper[9] Multi-scale object detection and accuracy in complex environments, sensitivity to
category has grown to be important duties in lighting and weather conditions, and high
numerous domain names, including laptop installation and maintenance costs. Conventional
imaginative and prescient, autonomous driving, image processing techniques struggle to handle
and surveillance. The task aims to develop a occlusions, varying vehicle sizes, and dynamic
complete and green approach for multi-scale backgrounds, leading to misclassifications and
item detection and classification with the usage false detections. Moreover, rule-based and
of machine cutting-edge and image processing thresholding methods lack adaptability, making
strategies. The proposed framework contains them inefficient for real-time traffic analysis.
three principal components - object detection, Due to these constraints, there is a growing need
function extraction, and type. for advanced machine learning-based solutions
that can enhance the accuracy and robustness of
In this paper[10] Artificial intelligence (AI) in object detection in traffic monitoring systems.
computer science is a branch that focuses on
creating intelligent systems or robots capable of
3.2 PROPOSED SYSTEM
imitating human behavior and reactions.
Artificial intelligence is a branch of computer The proposed system integrates machine
science. The impressive ability of individuals to learning (ML) and image processing to enhance
easily identify and differentiate items using their object detection in traffic cameras, providing

Page 8 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 9 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

higher accuracy, adaptability, and real-time

analysis. Unlike traditional methods, this system
leverages deep learning models such as
Convolutional Neural Networks (CNNs), You
Only Look Once (YOLO), and Faster R-CNN for
efficient and precise detection of vehicles,
pedestrians, and traffic elements. Advanced ML
models improve detection accuracy by
recognizing various vehicle types, pedestrians,
and road signs in real-time. The system utilizes
CNNs to extract meaningful visual features,
enabling better classification and recognition
under varying lighting and weather conditions.
Optimized algorithms ensure quick detection and
tracking, reducing response time for traffic
management and accident prevention.

3.3 ALGORITHM
Convolutional Neural Networks (CNNs)
algorithm The primary purpose of Convolution
in case of a CNN is to extract features from the Fig (a) CNN Models
CNN Layers
input image. Convolution preserves the spatial
• Convolutional Layer: Extracts features
relationship between pixels by learning image like edges, corners, textures.
features using small squares of input data. The • Activation Layer (ReLU): Adds non-
convolution layer’s parameters consist of a set of linearity and enhances feature learning.
• Pooling Layer: Reduces the spatial size
learnable filters. Every filter is small spatially
of feature maps.
(along width and height), but extends through the • Flattening: Converts 2D feature maps
full depth of the input volume. During the into a 1D vector.
forward pass, each filter is convolved across the • Fully Connected Layer: Learns high-
level representations and makes
width and height of the input volume and predictions.
computes dot products between the entries of the • Output Layer: Produces the final result
filter and the input at any position. As the filter (e.g., class probabilities).
• Backpropagation: Adjusts weights to
convolves over the width and height of the input
minimize the loss function.
volume it produces a 3-dimensional activation Convolution Operation
map that gives the responses of that filter at every The core of a CNN is the convolution operation.
spatial position. Convolution applies a filter (or kernel) to the
input image (or feature map) to extract features
like edges or textures.

Page 9 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 10 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

map, such as 13*13, the input image is divided

into 13*13 grids. The center of the object label
box is in a grid unit, and the grid unit is
responsible for predicting the object. The
network structure adopted by CNN is called
Darknet-53. This structure adopts the full
convolution method and replaces the previous
version of the direct-connected CNN with the
residual structure
Joining of the advanced innovations in the
observation of metropolitan versatility, for
3.4ARCHITECTURE DIAGRAM example, checking the traffic thickness,
Image
detecting the person wear a helmet will help in
Video Camer
segmentatio
working on the amount of vehicles/plans to be
a
n
accommodated the public compensation, office
capturi
to be fused in decreasing the traffic, foundation
to be given. For example, street extending,
person on foot way, without helmet, over span,
underpass and so on, where traffic and transport
is an issue. This can be executed in the city, at
OBJECT which it is perceived to be created as a savvy city.
DETECTIO
N CNN The proposed research work dissects the vehicle
ALGORITH thickness and detecting helmets utilizing python
M OpenCV and CNN algorithm. The foundation
extraction technique is applied for each casing
which is utilized in ensuing investigation to
Vehicle recognize and count every one of the vehicles.
detection
and A. DATASET
counting
Dataset : www.kaggle .com
Count no of
vehicles This is traffic camera that contains both vehicle
.I collected this dataset for my research work on
the traffic camera and mainly focused on no of
vehicle counting datasets specifically for traffic
4. EXPERIMENTAL RESULTS vehicle counting, there are several publicly
available datasets that can be used to train
This section describes the object models for vehicle detection and counting in
detection methods used in this study. The traffic scenes. These datasets often include
implementation of the highway vehicle detection various vehicle types (cars, trucks, buses,
CNN network. The convolutional neural bicycles, etc.), under different conditions, such
network is used to extract the features of the as varying weather, lighting, and traffic
input image. According to the size of the feature congestion.

Page 10 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 11 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

Video Dataset preprocessing frameworks for vehicle counting

and detection. First, resize the frames to be
consistent to ensure consistency across the
dataset and reduce computational complexity
during training. Then, convert the frames to a
standard color, such as RGB or grayscale, to
simplify the process and reduce noise. Use
techniques such as histogram equalization or
gamma correction to improve the contrast and
visibility of objects in frame.
DETECTING MOVING OBJECT
The implementation of the highway
vehicle detection framework used the CNN . The
CNN algorithm continues the basic idea of the
first two generations of Convolution neural
network algorithms. The CNN algorithm is used
to extract the features of the input image.
According to the size of the feature map, such as
13*13, the input image is divided into 13*13
grids. The centre of the object label box is in a
grid unit, and the grid unit is responsible for
predicting the object.
VEHICLE COUNTING
Vehicle counting entails detecting and
monitoring vehicles in a video feed to determine
the overall range of vehicles passing thru a
specific region or along a specific course. This
manner commonly includes several steps: First,
detecting shifting items within the video frames,
usually completed via heritage subtraction or
item detection strategies.

B.IMPLEMENTATION MODULE
DISPLAY RESULT
ACQUIRE VIDEO STREAM OR IMAGES Displaying the results of a vehicle
Datasets collected to count and detect counting system involves presenting the
vehicles Take pictures or videos containing information gleaned from the analysis of video
various traffic information such as highways, frames in a comprehensible format. This can be
intersections, or parking lots and use tick boxes accomplished by visually rendering the output on
or labels to indicate vehicle location and location a graphical user interface or screen.
which contains writing about. Result Analysis Main Page
PREPROCESS THE FRAMES
There are several steps to improve the
quality of the data and make it usable in

Page 11 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 12 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

C.RESULT

Fig (D) : Object Detection Vehicle Counting

Fig (A) : Object Detection Vehicle`

Fig (E) : Object Detection Vehicle Counting

Result

4.4 PERFORMANCE ANALYSIS

Performed on the diabetes dataset. Machine with

configuration of windows 10 system and 8-GB
Fig (B) : Pre-process CNN Result of RAM is used. This is traffic camera that
contains both vehicle .I collected this dataset for
my research work on the traffic camera and
mainly focused on no of vehicle counting
datasets specifically for traffic vehicle counting,
there are several publicly available datasets that
can be used to train models for vehicle detection
and counting in traffic scenes. These datasets
often include various vehicle types (cars, trucks,
buses, bicycles, etc.), under different conditions,
such as varying weather, lighting, and traffic
congestion.The results were compared to
experiments with Python implementations of
Fig (C) : Object Detection Vehicle Counting

Page 12 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 13 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

CNN,RF,SVM the techniques run to ensure that presents a viable solution for real-world
the results are comparable for accuracy. applications such as traffic monitoring and
surveillance. The algorithm demonstrates
Table 4.4 Traffic Dataset
effectiveness in detecting vehicles and accurately
Comparison of dataset using SVM, RF,ORB,CNN counting their movements within monitored
areas. While it shows promise, there are areas for
Detection False improvement, particularly in handling
Accuracy alarm occlusions and crowded traffic scenarios. Future
Algorithms Rate efforts should focus on refining the algorithm's
performance and optimizing its efficiency to
SVM 79% 0.61%
ensure robust operation in diverse environmental
RANDOM 79.6% 0.59% conditions. Overall, the CNN algorithm provides
FOREST a solid foundation for further research and
development in the field of computer vision and
ORB 81.9% 0.73%
intelligent transportation systems.`
CNN 99.8% 1.35%
REFERENCES

1. Venkata Naresh Mandhala, Debnath

Bhattacharyya, Vamsi B, Thirupathi Rao N,
“Object Detection Using Machine Learning
for Visually Impaired People” Scopus access,
published on 27 October 2020
2. Selim S. Sarikan, A. Murat Ozbayoglu,
“Anomaly Detection in Vehicle Traffic with
Image Processing And Machine Learning ”,
published on 7 November 2018 ,Access on
Fig (a): Accuracy ScienceDirect,Procedia Computer
0.016 140(2018)64-69
0.014 3. Mao Queen Garzón, “Inductive Machine
0.012 Learning with Image Processing for
0.01 ObjectsDetection of a Robotic Arm with
0.008
Raspberry PI” (August 2018), this
0.006
publication
0.004
0.002
athttps://www.researchgate.net/publication/3
0 3018320
1 2 3 4 5 4. Vandit Patela , Manthan Shaha , Shyam
Algorithms False alarm Rate Modia , Yash Patela , Sejal Thakkarb,
“Object Detection using Machine Learning”,
Fig (b) False Alarm Rate Object Detection Using Machine Learning: A
Survey published on 2021.
V.CONCLUSION
5. Chisulo Mukabe, Nalina Suresh, Valerians
In conclusion, the implementation of vehicle Hashiyana, Titus Haiduwa, William
detection and counting using the CNN algorithm Sverdlik,“Object Detection and

Page 13 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

 Page 14 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245

Classification Using Machine Learning 13. Kiran, V. K., Parida, P., & Dash, S. (2020,
Techniques”, A Comparison of Haar August). Vehicle detection and classification:
Cascades and Neural Networks, August 9-12, a
2021 14. review. In International Conference on
6. Prabakar Dakshinamoorthy , Innovations in Bio-Inspired Computing and
Gnanajeyaraman Rajaram ,Shruti garg Applications (pp. 45-56).
,Prabhu Murugan A. Manimaran ,Ramesh 15. Swathy, M., Nirmala, P., & Geethu, P. (2017).
Sundar ,” Arti¦cial Intelligence Algorithms Survey on vehicle detection and tracking
For Object Detection and Recognition In techniques in video surveillance.
video and Images”. Published on International Journal of Computer
january12th,2024. Applications, 160(7), 22-25.
7. Ganesh, B. S. V., & Kumar, N. S. (2024, 16. Hou, Y., Li, Q., Zhang, C., Lu, G., Ye, Z.,
July). A novel approach to machine learning Chen, Y., ... & Cao, D. (2021). The state-of-
for object detection and recognition. In the- art review on applications of intrusive
International Conference on Computational sensing, image processing techniques, and
Innovations and Emerging Trends (ICCIET- machine learning methods in pavement
2024) (pp. 918-925). Atlantis Press. monitoring and analysis. Engineering, 7(6),
8. Kumar, S., Sinha, A. K., Kumar, N., Kashif, 845-856.
M., & Hussain, A. (2020). Literature Review 17. Fredianelli, L., Carpita, S., Bernardini, M.,
of Moving Object Detection Using Machine Del Pizzo, L. G., Brocchi, F., Bianco, F., &
Learning. Journal of Intelligent & Licitra, G. (2022). Traffic flow detection
FuzzySystems, 39(2), 2761-2771. using camera images and machine learning
9. Mahmood, S. S., & Saud, L. J. (2020). An methods in ITS for noise map and action plan
efficient approach for detecting and optimization. Sensors, 22(5), 1929.
classifying moving vehicles in a video based 18. Khan, H., Kushwah, K. K., Maurya, M. R.,
monitoring system. Engineering and Singh, S., Jha, P., Mahobia, S. K., ... &
TechnologyJournal, 38(6A), 832-845. Sadasivuni, K. K. (2022). Machine learning
10. Ferjani, I., & Alsaif, S. A. (2022). How to get driven intelligent and self adaptive system
best predictions for road monitoring using for traffic management in smart cities.
machine learning techniques. PeerJ computing, 1-15.
Computer Science, 8, e941. 19. Shamrat, F. J. M., Mahmud, I., Rahman, A.
11. Yuvaraj, N., Rajput, K., Suganyadevi, K., S., Majumder, A., Tasnim, Z., & Nobel, N.
Aeri, M., Shukla, R. P., & Gurjar, H. (2024, I.(2020). A smart automated system model
May). Multi-Scale Object Detection and for vehicles detection to maintain traffic by
Classification using Machine Learning and image processing. International Journal of
Image Processing. In 2024 Second Scientific & Technology Research, 9(02),
International Conference on Data Science 2921-2928.
and Information System (ICDSIS) (pp. 1-6). 20. Alghamdi, A. S., Saeed, A., Kamran, M.,
IEEE. Mursi, K. T., & Almukadi, W. S. (2023).
12. Morris, C., & Yang, J. J. (2021). A machine Vehicle classification using deep feature
learning model pipeline for detecting wet fusion and genetic algorithms. Electronics,
pavement condition from live scenes of 12(2), 280.
traffic cameras. Machine Learning with
Applications, 5, 100070.

Page 14 of 14 - Integrity Submission Submission ID trn:oid:::9380:83833245