Volume 9, Issue 6, June – 2024 International Journal of Innovative Science and Research Technology

ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24JUN1260

Vehicle Detection System Using Machine Learning

Sadheen Hossain1; Aous Shaheen2; Dr. B.S. Satpute3; Tariqul Islam Sani4
19BTRCS1301; 19BTRCS1212; Associate Professor 3; 19BTRCV0604
Department of CSE, Faculty of Engineering &
Technology Jain (Deemed-to-be University

Abstract:- This project focuses on developing a vehicle This research article aims to present a detailed account
detection system using OpenCV, a real-time computer of thevehicle detection project using OpenCV.
vision library in Python. The project aims to create a
vehicle counting and detection system that works II. LITERATURE REVIEW
effectively for videos using OpenCV for image processing.
The system will utilize OpenCV's computer vision  "Vehicle Detection and Tracking in Video Surveillance:
capabilities to identifyvehicles and count the number of A Review" by H. Khan et al. (2018) This literature review
vehicles along with adding a unique id for each vehicle in provides an overview of vehicle detection and tracking
the video. The system has potential applications in traffic techniques in video surveillance. The authors explore the
monitoring, parking management, and transportation various approaches used for vehicle detection and
planning. The results of the project demonstrate the tracking, including feature-based methods, background
capabilities of OpenCV in creating efficient and accurate subtraction,and deep learning-based approaches.
vehicle detection and classification systems.  "A Survey of Vehicle Detection and Tracking
Techniques" by S. Singh and S. Singh (2020) This
Keywords:- Vehicle Detection, Object Detection, Opencv, literature review provides a comprehensive survey of
Image Processing, Computer Vision. vehicle detection and tracking techniques, including
traditional techniques and deep learning-based
I. INTRODUCTION approaches. The authors evaluate the performance of
these techniques and discuss the challenges associated
Road accidents not only cause a tragic loss of human with each approach.
lives, but also lead to significant financial and infrastructure  "Vehicle Detection and Tracking: A Review" by R. P.
damage. According to a study conducted by Bosch, road Kumawat and N. Bhatia (2019) This literature review
accidents resulted in a loss of nearly $14 billion to $40 billion provides an overview of vehicle detection and tracking
in the year 2019. To address this problem, it is necessary to systems, including the various techniques used for feature
observe the movements of vehicles using advanced extraction, object detection, and object tracking. The
implementations of object detection systems. Vehicle authors discuss the challenges associated with vehicle
detection systems can help in overcoming the challenges of detection and tracking and highlight the potential
locating vehicles in images collected from a moving applications of such systems.
platform.  "Vehicle Detection and Tracking: A Review of Recent
Literature" by A. K. Jain et al. (2021) This literature
A vehicle detection system is an application of an object review provides a detailed analysis of recent developments
detection system that focuses on tracking vehicles using their in vehicle detection and tracking techniques. The authors
coordinates in images or videos. These systems can be explore the various approaches used for feature
implemented using various tools, and in this research article, extraction, object detection, and object tracking and
we will be implementing a project using the OpenCV library. evaluate the performance of these techniques on different
Our computer vision system will use a combination of object datasets.
detection and object tracking to identify cars in a video stream at  "A Comprehensive Review of Vehicle Detection and
different waypoints. We will leverage OpenCV Machine Tracking in Intelligent Transportation Systems" by M. S.
Learning and image processing techniques to develop a Khan et al. (2018) This literature review provides a
vehicle detection system that can improve road safety and comprehensive overview of vehicle detection and
prevent accidents. tracking techniques in intelligent transportation systems.
The authors explore the various approaches used for
Our proposed system has several potential applications, feature extraction, object detection, and object tracking,
such as traffic monitoring, parking management and and discuss the challenges associated with each approach.
transportation planning. By accurately detecting and tracking They also highlight the potential applications of vehicle
vehicles, our system can enhance road safety and mitigate the detection and tracking systems in intelligent
risk of accidents. We believe that the implementation of such transportation systems.
a system can led to significant improvements in road safety and
help reduce the number of road accidents worldwide.

IJISRT24JUN1260 www.ijisrt.com 2489

Volume 9, Issue 6, June – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24JUN1260

III. METHODOLOGY  Vehicle Counting and Classification: We will use the

tracked vehicle data to count the number of vehicles in the
 Data Collection: The first step in our methodology is to video and assign a unique ID to each vehicle for
collect a dataset of videos containing vehicles. We will classification purposes. The classification will help us
use publicly available datasets such as the Caltech dataset, distinguish between different types of vehicles, such as
which contains around 250,000 images of vehicles in cars,trucks, and buses.
different scenarios.  Implementation: Finally, we will implement our system
 Preprocessing: Before feeding the videos into our system, using Python and the OpenCV library. We will create a
we need to preprocess them to ensure uniformity in the user-friendly interface that allows users to input a video
videoframes. We will resize the frames to a standard size, file and view the results of our vehicle detection system in
convert them to grayscale, and apply normalization real time.
techniques to ensure that our system can work with
varying lighting conditions. IV. RESULTS
 Object Detection: The next step is to detect vehicles in the
video frames. We will use a pre-trained classifier to  After implementing and running the program using visual
identifythe presence of vehicles in each frame. studio, we used a pre-recorded video as input, and we
 Object Tracking: Once we have detected the presence of have the following results:
vehicles in each frame, we will track their movements.
Object tracking will help us identify the trajectory of each
vehicle and its speed.

Fig 1: Program output

In Fig-1, we can see that the program is working To keep tracking all vehicles that are in the video, the
properly, where the vehicle count can be shown as numbers program provides a tracking log for all vehicles, as in the
on each vehicle. following output photo:

IJISRT24JUN1260 www.ijisrt.com 2490

Volume 9, Issue 6, June – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24JUN1260

Fig 2: Tracking Log for the Moving Vehicles

And as the number of vehicles is changes, while the and provides a new count for each one as well as a new
video isrunning and vehicles are moving, and new vehicles trackinglog, as in the following results:
show upin the video, the program detects the new vehicles

Fig 3: Program Output

A new tracking log is done as new vehicles are there, as the following:

IJISRT24JUN1260 www.ijisrt.com 2491

Volume 9, Issue 6, June – 2024 International Journal of Innovative Science and Research Technology
ISSN No:-2456-2165 https://doi.org/10.38124/ijisrt/IJISRT24JUN1260

Fig 4: Tracking Log for the Moving Vehicles

V. CONCLUSION REFERENCES

In this project, we developed an advanced vehicle [1]. Bradski G., & Kaehler A. (2008). Learning OpenCV:
detection and classification system using OpenCV. Our Computer vision with the OpenCV library. O'Reilly
system iscapable of detecting and tracking vehicles in real- Media, Inc.
time video streams, estimating their speeds, and storing [2]. Li H., Li Y., Li B., & Zou Y. (2019). Vehicle detection
evidence in thecloud. We achieved this by combining object and tracking in a multi-camera system. Journal of
detection and object tracking techniques, which allowed us to Visual Communication and Image Representation.
accurately detect and track vehicles in various lighting and [3]. Khan F. S., Anwer R. M., & van de Weijer J. (2016).
weather conditions. Object detection in videos with tubelet proposal
networks. In Proceedings of the IEEE Conference on
Our research contributes to the growing body of Computer Vision and Pattern Recognition.
literature onvehicle detection and tracking systems, providing [4]. Guo Y., Liu Y., Ouyang W., & Li X. (2019). Deep
a detailedanalysis of the techniques used in our system and learning for pedestrian detection: A comprehensive
their potential applications. Our results demonstrate the review. Neurocomputing.
effectiveness of our approach and highlight the importance of [5]. Ghosal S., Verma Y., & Mehrotra S. (2018). A survey
implementing such systems for improving road safety and on object detection and tracking in video surveillance.
reducing the economic and human costs associated with road Journal of Ambient Intelligence and Humanized
accidents. Computing.

Future research could explore further enhancements to

our system, such as improving the accuracy of speed
estimation and incorporating more advanced machine
learning algorithms for object detection and tracking. Overall,
our study emphasizes the value of utilizing computer vision
technologies to enhance road safety and reduce the impact of
road accidents on society.

IJISRT24JUN1260 www.ijisrt.com 2492