VISVESVARAYA TECHNOLOGICAL UNIVERSITY

Jnana Sangama, Belagavi -590018

A MINI PROJECT REPORT

ON
“OBJECT TRACKING WITH OPENCV”
Submitted in the partial fulfillment as the requirement of
V semester for the award of the degree of

BACHELOR OF ENGINEERING
IN
ARTIFICIAL INTELLIGENCE AND DATA SCIENCE
Submitted by

ADVIK 4YG22AD002

FAHAD AMEEN 4YG22AD017

MOHAMMED IMDAD KHAN H.J 4YG22AD031

SHREYAS RAJ 4YG22AD048

Under the Guidance of

Mrs. Varshitha B.E., M.Tech.,
Assistant Professor
Dept. of AI&DS, NCE, Hassan

DEPARTMENT OF ARTIFICIAL INTELLIGENCE

AND DATA SCIENCE

NAVKIS COLLEGE OF ENGINEERING, HASSAN

HASSAN-573217
2024-2025
 NAVKIS COLLEGE OF ENGINEERING HASSAN
(Affiliated to Visvesvaraya Technological University)

Department of Artificial Intelligence and Data Science

CERTIFICATE
This is to Certify that the mini project work entitled “OBJECT TRACKING WITH OPENCV”
is a bonafide work carried out by ADVIK(4YG22AD002), FAHAD AMEEN(4YG22AD017),
MOHAMMED IMDAD KHAN H.J(4YG22AD031), SHREYAS RAJ(4YG22AD048) in partial
fulfillment for the award of Bachelor of Engineering in Artificial Intelligence and Data Science
of the Visvesvaraya Technological University, Belagavi, Karnataka during the year 2024-2025.
It is certified that all corrections/suggestions indicated for the Internal Assessment have been
incorporated in the report. The mini project report has been approved as it satisfiesthe academic
requirements in respect of mini project work prescribed for the Bachelor of Engineering degree.

--------------------------- ------------------------------
Signature of Guide Signature of HOD
Mrs. Varshitha M Mr. Vivekananda
Assistant Professor Dept. of AI&DS
Dept. of AI&DS NCE.,Hassan
NCE.,Hassan
 DECLARATION

We, the under signed students of 5th semester Artificial Intelligence and Data Science, Navkis
College of Engineering, Hassan. Solemnly declare that our mini project work entitled
“OBJECT TRACKING WITH OPENCV” is a bonafide work of ours. Our project is neither
a copy nor by means a modification of any other engineering project.
We also declare that this project was not entitled for submission to any other university in the
past and shall remain the only submission made and will not be submitted by us to any other
university in the future.

Name USN Signature

ADVIK (4YG22AD002)

FAHAD AMEEN (4YG22AD017)

MOHAMMED IMDAD KHAN (4YG22AD031)

H. J
SHREYAS RAJ (4YG22AD048)

Place: Hassan
Date:
 ACKNOWLEDGEMENT

It is a great pleasure for us to acknowledge the help of many individuals without the help of
those this project would not have been fruitful.

We take this opportunity to express our sincere gratitude and respect to the Navkis College of
Engineering, Hassan for providing us an opportunity to carry out project.

We are ardently thankful to Dr. Venu Gopal Rao, Principal, Navkis College of Engineering,
Hassan for their encouragement and useful suggestions for carrying out this successfully.

We acknowledge with a deep sense of obligation, the encouragement given by Head of the
department, Department of Artificial Intelligence and Data Science, Navkis College of
Engineering, Hassan.

We greatly acknowledge our guide Mrs. Varshitha M, Assistant Professor, Department of

Artificial Intelligence and Data Science, Navkis College of Engineering, Hassan for providing
therequired impet us and helping us for carrying out this project successfully.

We respectfully thank all the Faculty Members and Staffs of the Department of Artificial
Intelligence and Data Science, Navkis College of Engineering, Hassan who had us in the
completion of this project.

We would like to thank all our Friends, who helped us directly or indirectly in the successful
completion of this project. Lastly but most importantly, we are thankful to our beloved Parents
and other Family members, without whose blessings and constant inspiration this project
would not have been a success.

ADVIK 4YG22AD002
FAHAD AMEEN 4YG22AD017
MOHAMMED IMDAD KHAN H.J 4YG22AD031
SHREYAS RAJ 4YG22AD048
 ABSTRACT

Object tracking is a fundamental task in the field of computer vision, with a wide range of
applications, including video surveillance, robotics, and sports analytics. In this study, we
present a robust and efficient object tracking system developed using the OpenCV library, a
popular open-source computer vision and machine learning software library. The proposed
approach combines advanced computer vision techniques to accurately track the movement of
an object in real-time. The system first identifies the target object based on its color
characteristics, utilizing the HSV color space to create a precise color mask. Next, contour
detection is employed to accurately locate the object's position within each frame. To enhance
the tracking stability and handle potential occlusions, a Kalman filter is integrated to predict
the object's future position, allowing the system to maintain a smooth and continuous tracking
even when the object is temporarily obscured from view
Contents:
CHAPTER1: INTRODUCTION

1.1: Introduction about the project

1.1.1 : Motivation

1.1.2 : Problem statement

1.1.3 : Scope of the project

1.1.4 : Objectives

1.2: Review of literature

CHAPTER 2: SYSTEM REQUIREMENT SPECIFICATION

2.1: Specific requirements

2.2: Hardware requirements

2.3: Software requirements

CHAPTER 3: DETAILED DESIGN

3.1: Design considerations

3.2: System architecture

3.2: Module specification

3.4: Data Flow Diagram

CHAPTER 4: IMPLEMENTAION

4.1: Implementation requirements

4.1.1: Programming language selection

4.1.2: Key features of programming language selected

4.2: Coding guidelines for programming language used in the project

4.3: Pseudo code for each module with description

CHAPTER 5: SYSTEM TESTING

 5.1: Test procedures

5.2: Unit test cases

5.3: Integrated test for the system

CHAPTER 6: RESULTS AND ANALYSIS

6.1: Results /Snapshots with description of each module

CHAPTER 7: CONCLUSION AND FUTURE ENHANCEMENTS

7.1: Conclusions

7.2: Future enhancements

CHAPTER 8: APPENDIX

8.1: Results and Discussions

REFERENCES
 OBJECT TRACKING WITH OPENCV 2024-25

CHAPTER 1
INTRODUCTION

1.1 INTRODUCTION OF THE PROJECT

Computer vision and object tracking are rapidly growing fields with a wide range of real-world
applications, from autonomous vehicles and surveillance systems to human-computer interaction and
sports analytics. One of the fundamental tasks in computer vision is the ability to detect and track the
movement of objects within a video stream or sequence of images. This project explores the
implementation of object tracking using the popular OpenCV computer vision library, with a ball serving
as the example object.

OpenCV, short for Open Source Computer Vision Library, is an open-source computer vision and
machine learning software library that has been widely adopted by researchers, developers, and
enthusiasts alike. Its extensive set of functions and algorithms make it a powerful tool for a variety of
computer vision tasks, including object detection, image processing, and video analysis. By leveraging
the capabilities of OpenCV, this project aims to demonstrate the process of tracking the movement of a
ball in real-time, showcasing the versatility and practicality of this technology.

The ability to accurately track the motion of objects has numerous applications in fields such as sports
analytics, robotics, surveillance, and augmented reality. In the context of sports, object tracking can be
used to analyze the trajectories of balls or other sports equipment, providing valuable insights for coaches,
players, and analysts. In robotics, object tracking is essential for tasks like obstacle avoidance and
autonomous navigation. Similarly, surveillance systems rely on object tracking to monitor and analyze
the movement of people and vehicles, while augmented reality applications utilize object tracking to
seamlessly overlay digital content onto the physical world

Dept. of AI&DS, NCE, Hassan Page:1

 OBJECT TRACKING WITH OPENCV 2024-25

1.1.1. MOTIVATION

Object tracking is a fundamental task in computer vision with a wide range of real-world
applications, from autonomous vehicles and robotics to video surveillance and sports analytics.
Tracking the motion of objects as they move through a scene is crucial for understanding and
interpreting dynamic environments. In the context of this project, we have chosen to focus on
tracking a ball as a specific example of object tracking using OpenCV, a popular open-source
computer vision library. The motivation behind this project is multifaceted. Firstly, ball tracking
has significant practical applications in sports analysis and coaching. By accurately tracking the
trajectory of a ball, coaches and analysts can gain valuable insights into player performance, team
strategies, and the dynamics of the game. This information can be used to optimize training,
improve decision-making, and enhance the overall quality of the sport.
Additionally, ball tracking serves as an excellent case study for the broader challenges and
techniques involved in object tracking. Balls, with their distinct shape, color, and motion patterns,
provide a well-defined and relatively simple object to track. By tackling the problem of ball
tracking, we can develop and refine the core algorithms and methodologies that underpin more
complex object tracking tasks.
Furthermore, the exploration of ball tracking using OpenCV presents an opportunity to showcase
the versatility and power of this open-source library. OpenCV has a rich set of computer vision
algorithms and tools that can be leveraged to address a wide range of problems, including object
detection, segmentation, and tracking. By demonstrating the effectiveness of OpenCV in ball
tracking, we can inspire and inform future researchers and developers who seek to apply computer
vision techniques to their own areas of interest.
In summary, the motivation for this project on Object Tracking with OpenCV, using a ball as the
specific example, is driven by the practical applications in sports analytics, the educational value
in exploring fundamental object tracking techniques, and the desire to showcase the capabilities of
the OpenCV library. Through this work, we aim to contribute to the advancement of computer
vision research and its real-world impact

Dept. of AI&DS, NCE, Hassan Page:2

 OBJECT TRACKING WITH OPENCV 2024-25

1.1.2 PROBLEM STATEMENT

The task of tracking a ball's motion in a dynamic environment presents a compelling challenge in
the field of computer vision. Accurately following the trajectory of a rapidly moving ball, while
accounting for factors such as varying lighting conditions, occlusions, and camera movements, requires
the development of sophisticated algorithms and techniques. Traditional approaches to ball tracking often
relied on specialized hardware, such as high-speed cameras and dedicated tracking sensors. While these
solutions can provide high accuracy tracking, they can also be prohibitively expensive and limited in
their adaptability to different scenarios. The emergence of open-source computer vision libraries, such
as OpenCV, has opened new avenues for more accessible and versatile ball tracking solutions. The
primary problem this project seeks to address is the development of a robust and efficient ball tracking
system using OpenCV.

The key objectives include:

• Object Detection: Accurately identifying the ball within a given frame, despite variations in
size, orientation, and environmental conditions.
• Object Tracking: Maintaining a continuous and reliable track of the ball's movement across
multiple frames, handling challenges like occlusions and sudden changes in direction.
• Trajectory Estimation: Reconstructing the ball's trajectory over time, providing valuable
insights into its path, velocity, and acceleration.
• Real-time Performance: Ensuring the tracking system can operate in real-time, processing
frames quickly enough to capture the dynamic nature of the ball's motion.
• Generalizability: Developing a solution that can be easily adapted to different ball-based
sports, camera setups, and environmental conditions, without requiring extensive manual
tuning or calibration.
Addressing these challenges requires a comprehensive understanding of computer vision techniques,
including object detection, feature extraction, and motion modeling.

Additionally, the integration of these techniques within the OpenCV framework presents unique
opportunities and constraints that must be carefully navigated. By successfully tackling this problem,
the project will contribute to the broader field of object tracking, showcasing the potential of open-
source computer vision tools to deliver cost-effective and versatile solutions. The insights gained from

Dept. of AI&DS, NCE, Hassan Page:3

 OBJECT TRACKING WITH OPENCV 2024-25

this work can inform future research and development in areas such as sports analytics, robotic vision,
and surveillance applications

1.1.3 SCOPE OF THE PROJECT

The project will encompass the following main elements:

• Ball Detection: Developing efficient algorithms for accurately detecting and localizing the ball
within each video frame. This will involve techniques such as color-based segmentation, shape-
based recognition, and machine learning-based object detection.
• Ball Tracking: Implementing robust tracking algorithms to maintain a continuous and reliable
trajectory of the ball's movement across multiple frames. This will include the use of techniques
like Kalman filtering, mean-shift tracking, and correlation-based methods.
• Trajectory Estimation: Reconstructing the ball's trajectory over time, providing detailed
information about its position, velocity, and acceleration. This will enable the analysis of key
performance metrics and strategic insights.
• Real-time Performance: Ensuring the developed ball tracking system can operate in real-
time, processing frames at a rate that can effectively capture the dynamic nature of the ball's
motion. Optimizing the algorithms and leveraging OpenCV's efficient computations will be
crucial.
• Generalization and Adaptability: Designing the tracking system to be easily adaptable to
different ball-based sports, camera setups, and environmental conditions. This will involve
incorporating techniques for automatic parameter tuning and handling of varying lighting,
occlusions, and camera movements.
• Evaluation and Benchmarking: Rigorously evaluating the performance of the ball tracking
system using established metrics, such as precision, recall, and F-score. Comparing the results
against state-of-the-art methods and benchmarking the system's capabilities.
• Practical Applications: Exploring the potential applications of the developed ball tracking
system, with a particular focus on sports analytics and coaching. Demonstrating how the
tracking data can be utilized to provide valuable insights and support decision-making
processes.
By addressing these aspects within the scope of the project, the goal is to create a comprehensive and
robust ball tracking solution that showcases the capabilities of OpenCV and serves as a valuable resource
for researchers, developers, and practitioners in the field of computer vision and sports analytics.

Dept. of AI&DS, NCE, Hassan Page:4

 OBJECT TRACKING WITH OPENCV 2024-25

1.1.4 OBJECTIVES

The primary objectives of this project on Object Tracking with OpenCV, using a ball as the specific
example, are as follows:
1. Accurate Ball Detection:-
Develop efficient algorithms for accurately detecting and localizing the ball within each video frame,
even in the presence of varying lighting conditions, occlusions, and camera movements. Explore a
range of object detection techniques, including color-based segmentation, shape-based recognition, and
machine learning-based object detection. Ensure the ball detection process is robust and can handle
diverse ball characteristics, such as size, color, and texture.

2.Reliable Ball Tracking:-

Implement robust tracking algorithms to maintain a continuous and reliable trajectory of the ball's
movement across multiple frames. Investigate the use of techniques like Kalman filtering, mean-shift
tracking, and correlation-based methods for ball tracking. Ensure the tracking system can effectively
handle challenges such as occlusions, sudden changes in direction, and loss of visual cues.

3.Trajectory Estimation and Analysis:-

Reconstruct the ball's trajectory over time, providing detailed information about its position, velocity,
and acceleration. Develop methods to extract and analyze key performance metrics and strategic
insights from the ball's trajectory data. Explore the application of the tracking system in sports
analytics, coaching, and decision-making processes.

4. Real-time Performance:-
Ensure the developed ball tracking system can operate in real-time, processing frames at a rate that
effectively captures the dynamic nature of the ball's motion. Optimize the algorithms and leverage
OpenCV's efficient computational capabilities to achieve high-performance tracking.

5. Generalization and Adaptability:-

Design the tracking system to be easily adaptable to different ball-based sports, camera setups, and
environmental conditions. Incorporate techniques for automatic parameter tuning and handling of
varying lighting, occlusions, and camera movements. Ensure the system's flexibility and extensibility to
support a wide range of applications beyond the specific ball tracking use case.

Dept. of AI&DS, NCE, Hassan Page:5

 OBJECT TRACKING WITH OPENCV 2024-25

6. Evaluation and Benchmarking:-

Rigorously evaluate the performance of the ball tracking system using established metrics, such as
precision, recall, and F-score. Compare the results against state-of-the-art methods and benchmark the
system's capabilities to assess its effectiveness and competitiveness.

By achieving these objectives, the project aims to deliver a comprehensive and robust ball tracking
solution that showcases the power of OpenCV in computer vision applications, particularly in the realm
of sports analytics and coaching.

1.2 REVIEW OF LITERATURE

The ability to accurately and reliably track the motion of objects, such as balls in various sports, has
become increasingly important in computer vision and sports analytics. Object tracking is a fundamental
task that enables a wide range of applications, including performance analysis, tactical decision-making,
and immersive viewing experiences. This literature review examines the state-of-the-art in object tracking
techniques, with a specific focus on ball tracking using the OpenCV computer vision library.

[1] “Object Tracking: A Survey” by Alper Yilmaz, Omar Javed, and Mubarak Shah (2006)

Advantages: Provides a comprehensive overview of object tracking techniques, including feature-based,

model-based, and adaptive methods.
Discusses the challenges and future directions in object tracking, which are still relevant today.
Disadvantages and Success Rate: Does not specifically focus on ball tracking or the use of OpenCV.
Covers a broad range of object tracking techniques, without in-depth analysis of any specific approach.
This survey paper has been highly influential in the field of object tracking, with over 5,500 citations. It
serves as a valuable reference for researchers and developers working on object tracking problems.

[2] "Learning OpenCV: Computer Vision with the OpenCV Library" by Gary Bradski and
Adrian Kaehler (2008)

Advantages: Provides a comprehensive introduction to the OpenCV library and its various computer
vision algorithms, including object tracking techniques.
Covers the fundamentals of OpenCV programming and how to apply it to real-world problems.

Dept. of AI&DS, NCE, Hassan Page:6

 OBJECT TRACKING WITH OPENCV 2024-25

Disadvantages and Success Rate: The book is relatively old, and some of the information may not be
up-to-date with the latest developments in OpenCV.
Does not focus specifically on ball tracking or advanced object tracking techniques.
This book is considered a classic in the field of computer vision and OpenCV programming. It has been
widely used by students, researchers, and practitioners as a valuable resource for learning and applying
OpenCV.

[3]"Simple Online and Realtime Tracking" by Alex Bewley, Zongyuan Ge, Lionel Ott, Fabio
Ramos, and Ben Upcroft (2016)

Advantages: Presents a simple, yet effective, online and real-time object tracking algorithm that can be
used for a variety of applications, including ball tracking. The algorithm is efficient and can run at high
frame rates, making it suitable for real-time tracking tasks.
Disadvantages and Success Rate :The paper focuses on the general object tracking problem and does
not provide specific insights into ball tracking .The algorithm may not be as accurate or robust as more
complex tracking methods, especially in challenging scenarios. The Simple Online and Realtime
Tracking (SORT) algorithm proposed in this paper has been widely adopted and used in various object
tracking applications. It has become a popular baseline for comparison in the object tracking community.

[4]"Mastering OpenCV 4 with Python" by David Millán Escrivá, Vinícius Godoy, and Michael
Beyeler (2018)

Advantages :Provides a comprehensive guide to using OpenCV 4 in Python, covering a wide range of
computer vision tasks, including object tracking. Includes practical examples and code snippets for
implementing various object tracking algorithms in OpenCV.
Disadvantages and Success Rate: While the book covers object tracking in general, it does not delve
deeply into specific techniques for ball tracking or sports-related application. The information may not
be as cutting-edge as the latest research papers in the field. This book is a valuable resource for
developers and researchers who want to learn how to use OpenCV effectively in their projects. It has
been well-received by the community and is considered a go-to reference for OpenCV programming in

Dept. of AI&DS, NCE, Hassan Page:7

 OBJECT TRACKING WITH OPENCV 2024-25

Python.

[5]"Deep Learning for Computer Vision with Python" by Adrian Rosebrock (2018)

Advantages: Provides a comprehensive introduction to deep learning techniques for computer vision
tasks, including object detection and tracking . Covers the latest advancements in deep learning-based
object tracking, which can be applicable to ball tracking scenarios.
Disadvantages and Success Rate: While the book discusses object tracking in general, it does not focus
specifically on ball tracking or the use of OpenCV.The content may be more oriented towards deep
learning practitioners rather than traditional computer vision researchers. Adrian Rosebrock's books and
tutorials on the PyImageSearch blog have gained a large following in the computer vision community.
This book is considered a valuable resource for practitioners looking to apply deep learning to solve
computer vision problems.

[6]"A Survey of Visual Object Tracking for Autonomous Vehicles" by Bing Wang, Guoliang Ye,
Shuai Li, and Xiaodong Liu (2019)

Advantages: Provides a comprehensive survey of object tracking techniques specifically tailored for
autonomous vehicle applications, which can be relevant to ball tracking in sports. Discusses the
challenges and requirements of object tracking in dynamic, real-world environments, which can inform
the design of ball tracking systems.
Disadvantages and Success Rate: While the survey covers object tracking in general, it does not focus
specifically on ball tracking or the use of OpenCV. The survey may be more oriented towards the
autonomous vehicle domain rather than sports analytics applications. This survey paper has been cited
over 100 times, indicating its relevance and impact in the field of object tracking for autonomous
vehicles. The insights it provides can be valuable for researchers and developers working on similar
problems in the sports domain.

Dept. of AI&DS, NCE, Hassan Page:8