EASWARI ENGINEERING COLLEGE

(AUTONOMOUS)
Bharathi Salai, Ramapuram, Chennai – 600 089

DEPARTMENT OF CHEMISTRY
[

SYNTHESIS AND CHARACTERIZATION OF CERTAIN

NOVEL TRANSITION METAL SUBSTITUTED
NANOCOMPOSITES AND ITS APPLICATION

A Minor project

Submitted by
1. Ajay 4. Chelsia
2. Akash 5. Dhanush
3. Balaji 6. Harini

Year: I Semester: II

Brach: Computer science and Engineering Section: A

April 2025
 EASWARI ENGINEERING COLLEGE
(AUTONOMOUS)
Bharathi Salai, Ramapuram, Chennai – 600 089

DEPARTMENT OF CHEMISTRY

CERTIFICATE

The Minor project work entitled “Synthesis and characterization of certain novel

transition metal substituted Nanocomposites and its application” has been Certified

by Department of Chemistry as a part of Internal assessment of Environmental science

for Computing sciences (231GEB201T) in the second semester of first year.

Name of the student Staff Incharge

XXXXXX
1. 4.
2. 5.
3. 6.

iii
 ABSTRACT

In this modern world, people have been living a life based on technology. The
progression of human society has developed most of its aspects into new standards.
Technologies have leapt through the times at a tremendous pace into development. As the
time passes new advancement in technology occurs, which makes our tasks easier and
efficient.

Smart healthcare system is an IoT based system which could be used for covid patients
for the current scenario and elderly people at our homes whose real time health readings
such as temperature, oxygen saturation level and heart rates could be monitored remotely
on a hand held device. This IoT device will automatically send an alert to the patient
intimates in case of an emergency which in this case would be fluctuation of the readings
of the sensors beyond the normal range. This device is being built using a temperature
sensor, electro-cardiogram sensor, gas sensor, pulse oximeter sensor attached to an
arduino mega board which will be transferring its data to the server using esp8266 wifi
module. The server will compute the data which could then be displayed on handheld
devices. In case the values received from the sensors are outside the normal range then an
alert will be sent to the user from the server in the form of a call alert and email. Patients
might be unable to make a call to the people around them in certain emergencies and in
this pandemic it is more prone to get exposed to viruses while monitoring the patient
being physically present, so from this project we are trying to deliver higher quality care
to more patients with a lower risk of exposure which can be helpful for the patients,
intimates, doctors and required authorities by continuously monitoring the patient’s status
and informing them in no time if any bad situation arises.

Keywords: oxygen-saturation level, temperature, monitor, handheld devices, exposure

iv
 TABLE OF CONTENTS

Title Page

CERTIFICATE ii
ABSTRACT iii
LIST OF FIGURES vi
LIST OF ABBREVIATIONS/SYMBOLS vii
CHAPTER 1 1
INTRODUCTION 1
1.1. Background 1
1.2. Motivation 1
1.3. Statement of the problem 2
1.4. Project objectives 2
1.5. Significance of the study 2
1.6. Applications 3
1.7. Features 3
1.8. Feasibility study 3
1.8.1. Operational feasibility 3
1.8.2. Economic feasibility 3
1.8.3. Technical feasibility 4
1.8.4. Schedule feasibility 4
CHAPTER 2 5
LITERATURE REVIEW 5
2.1. Existing prototypes 5
CHAPTER 3 8
REQUIREMENT ANALYSIS 8
3.1. Hardware Requirements 8

v
 3.1.1. Arduino Mega 2560 8
3.1.2. MAX 30100 Pulse Oximeter Sensor 8
3.1.3. MQ-2 gas sensor 10
3.1.4. LM35 temperature sensor 11
3.1.5. ECG AD8232 12
3.1.6. SIM 800A wireless GSM/GPRS module 12
3.1.7. ESP 8266 wifi module 13
3.2. Software Requirements 15
3.2.1. Arduino IDE 15
3.2.2. Visual Studio Code 16
3.2.3. MySQLi 16
3.2.4. IOT Cloud 16
3.2.5. Lucid chart 17
CHAPTER 4 18
SYSTEM DESIGN AND ARCHITECTURE 18
CHAPTER 5 19
METHODOLOGY 19
5.1. System Overview 19
5.2. System Flowchart 20
5.3. Dataflow Model 21
CHAPTER 6 22
RESULTS AND ANALYSIS 22
6.1. Output 22
6.2. Limitations 28
6.3. Future scopes 28
6.4. Work schedule 29
CHAPTER 7 30
TOTAL COST 30
CONCLUSION 31
REFERENCES 32

vi
 LIST OF FIGURES

Figure 3.1: Arduino mega board ------------------------------------------------------------------- 8

Figure 3.2: MAX 30100 pulse oximeter Sensor ------------------------------------------------- 9
Figure 3.3: MQ-2 gas sensor --------------------------------------------------------------------- 10
Figure 3.4: LM35 temperature sensor ------------------------------------------------------------ 11
Figure 3.5: AD8232 ECG sensor ------------------------------------------------------------------12
Figure 3.6: SIM 800A wireless GSM/GPRS module----------------------------------------- 13
Figure 3.7: ESP8266 wiFi module --------------------------------------------------------------- 14
Figure 3.8: Screenshot of Arduino IDE --------------------------------------------------------- 15
Figure 3.9: Screenshot of VS-code----------------------------------------------------------------16
Figure 4.1: Basic system overview diagram ----------------------------------------------------- 18
Figure 5.1: System flowchart ----------------------------------------------------------------------20
Figure 5.2: Dataflow model ----------------------------------------------------------------------- 21
Figure 6.1: Real time sensors value displayed on website ------------------------------------ 22
Figure 6.2: Database storing sensors value ------------------------------------------------------23
Figure 6.3: Values being displayed graphically on admin panel ------------------------------24
Figure 6.4: Call alert notifications ---------------------------------------------------------------- 25
Figure 6.5: Email alert notifications -------------------------------------------------------------- 26
Figure 6.6: Results on LCD ------------------------------------------------------------------------26
Figure 6.7: System setup ---------------------------------------------------------------------------27
Figure 6.8: System setup on the patient body --------------------------------------------------- 28

vii
LIST OF ABBREVIATIONS/SYMBOLS

ECG Electrocardiogram
API Application Programming Interface
PHP Hypertext Preprocessor
SQL Structured Query Language
VS Visual Studio
IoT Internet of Things
LCD Liquid Crystal Display
GSM Global System for Mobile Communication
GPRS General Packet Radio Service

viii
CHAPTER 1

INTRODUCTION

1.1. Background

The global plastic waste crisis has caused severe environmental degradation, with
traditional recycling methods failing to meet growing demands. Recent advancements in
computer vision (CV) and machine learning (ML) present opportunities to revolutionize
waste management practices. This project introduces an automated system to address
inefficiencies in plastic segregation and promote sustainable waste management.

1.2. Motivation

The motivation for this project stems from the pressing need to combat the challenges
posed by improper plastic waste segregation. Current manual processes are labor-
intensive, error-prone, and cost-ineffective. Leveraging AI technologies provides a
pathway to enhance recycling processes, alleviate environmental impacts, and contribute
to global sustainability efforts.

1
1.3. Statement of the problem

Manual and semi-automated plastic segregation methods are inefficient and result in
significant environmental and economic repercussions. There is an urgent need for an
automated system that can accurately and efficiently classify and separate plastics to
mitigate these issues and advance recycling efforts.

1.4. Project objectives

Main objective
● Develop a computer vision and machine learning-based system to classify and
segregate plastics automatically.
● Enhance recycling efficiency through real-time, accurate processing of waste
materials.
● Reduce environmental pollution caused by improper plastic waste management.
● Demonstrate the integration of AI-driven systems in industrial applications.

1.5. Significance of the study

This project highlights the potential of AI technologies in addressing critical

environmental issues. By automating plastic waste segregation, the study aims to
contribute to:
• Improved recycling rates.
• Reduced landfill deposits
• Enhanced resource recovery and environmental conservation.

2
1.6. Applications

● Recycling facilities to automate sorting.

● Waste management plants for efficient segregation
● Industrial settings like packaging industries to streamline waste disposal.
● Environmental cleanup initiatives targeting plastic waste in ecosystems.

1.7. Features

● Real-time Processing: Immediate classification and segregation.

● High Accuracy: AI-driven algorithms ensure precise identification

● Scalability: Adaptable for various industrial settings and waste streams.

● Environmentally Friendly: Contributes to sustainable waste management practices.

1.8. Feasibility study

1.8.1. Operational feasibility

The system is designed to function seamlessly in waste management environments,

utilizing conveyor belts, sensors, and sorting mechanisms to handle various plastic types
efficiently.

1.8.2. Economic feasibility

The initial cost of implementation is outweighed by long-term savings in labor, increased

recycling efficiency, and reduced environmental penalties.

3
1.8.3. Technical feasibility

Using advanced technologies like TensorFlow, OpenCV, and Python ensures the
technical capabilities to process data in real-time and integrate seamlessly with existing
waste management systems.

1.8.4. Schedule feasibility

The project timeline accounts for algorithm development, system design, and testing
phases, ensuring timely deployment within operational constraints.

4
CHAPTER 2

LITERATURE REVIEW

The field of plastic waste management has seen various approaches in addressing the
inefficiencies of traditional methods. Manual sorting systems are widely employed but
are labor-intensive, error-prone, and costly, making them unsuitable for large-scale
operations. Semi-automated methods offer slight improvements but still lack precision
and scalability. Advanced chemical recycling techniques, such as pyrolysis and
gasification, provide potential environmental benefits but involve high energy
consumption and costs. Recent developments in computer vision (CV) and machine
learning (ML) have facilitated the automation of waste segregation. Systems employing
image recognition algorithms and robotic sorting mechanisms have demonstrated
improved accuracy and efficiency. However, these systems often face challenges with
contaminated or complex waste streams and require extensive computational resources.

2.1. Existing prototypes

The journey of addressing plastic waste management has seen the evolution of various
prototypes, each offering incremental advancements. Traditional manual sorting systems
have been widely utilized due to their simplicity. However, these systems rely heavily on
human labor, making them prone to errors and inefficiencies. The inability to scale
effectively and the limitations in handling mixed or complex waste streams have made
manual methods less favorable for modern waste management needs.

Semi-automated systems emerged as an improvement over manual methods,

incorporating mechanical sorting mechanisms. While these systems reduce some
dependency on labor and enhance sorting accuracy to an extent, they still fail to meet the
efficiency and scalability required for industrial applications. Furthermore, semi-
automated systems often struggle with the rapid classification of diverse types of plastics.
With advancements in technology, computer vision-based prototypes began to address
5
some of these challenges. Utilizing image recognition algorithms, these systems classify
plastics based on their visual characteristics. Despite promising results, their reliance on
smaller datasets and simpler models often limits their accuracy when dealing with
contaminated or mixed plastic waste. As a result, their real-world applicability remains
constrained.

Another innovative approach involves robotic sorting systems, where robotic arms
equipped with basic classification mechanisms handle plastic segregation. These systems
show potential for improving efficiency, especially in controlled environments. However,
they face challenges in dealing with more complex and contaminated waste streams,
highlighting the need for further refinement.

Chemical and optical sorting prototypes, such as those using near-infrared spectroscopy,
focus on identifying plastics based on their chemical composition. While these methods
demonstrate high levels of precision, their high operational costs make them impractical
for widespread adoption. Recent developments in machine learning-enhanced systems
provide hope, as they leverage powerful algorithms to achieve higher accuracy in sorting.
However, the extensive computational power and large datasets required for these
systems represent significant barriers, emphasizing the need for cost-effective and
scalable solutions.

6
CHAPTER 3

REQUIREMENT ANALYSIS

3.1. Hardware Requirements

3.1.1. Arduino Mega 2560

The Arduino Mega 2560 is a microcontroller board based on the ATmega2560. It has 54
digital input/output pins (of which 15 can be used as PWM outputs), 16 analog inputs, 4
UARTs (hardware serial ports), a 16 MHz crystal oscillator, a USB connection, a power
jack, an ICSP header, and a reset button. It contains everything needed to support the
microcontroller; simply connect it to a computer with a USB cable or power it with a
AC-to-DC adapter or battery to get started. The Mega 2560 board is compatible with
most shields designed for the Uno and the former boards Duemilanove or Diecimila.

(Ref:https://www.circuitstoday.com/arduino-mega-pinout-schematics)
Figure 3.1: Arduino mega board

3.1.2. Servo Motor

A servo motor is an electrical device designed for precise control of angular or linear motion.
It uses a feedback mechanism to adjust its position, speed, and torque, making it ideal
for tasks that require accurate movements. In automation and robotics, servo motors are

7
widely used for their ability to perform specific movements repeatedly and efficiently.
They are compact, reliable, and versatile, making them suitable for various applications,
including sorting mechanisms in waste management systems. With its precision
and responsiveness, a servo motor ensures smooth and accurate operation in
systems requiring controlled motion.

(Ref: Servo Motor: Definition, Working Principle, and Applications | Electrical4U)

Figure 3.2: Servo motor

8
3.1.3. 4 Channel Module

A 4-channel relay module is an electronic device that allows you to control multiple high-
voltage devices using low-voltage signals. It consists of four relays integrated into a
single board, enabling the simultaneous operation of four independent circuits. Each relay
can switch devices like motors, lights, or appliances on and off, making it ideal for
automation projects..

Features:

● Can handle devices operating at up to 250VAC or 30VDC.

● Operates with a low input voltage, typically 5V, making it compatible with microcontrollers
like Arduino or Raspberry Pi.
● Provides electrical isolation between the control circuit and the high-voltage devices for
safety.
● Combines four relays into a single module, saving space and simplifying wiring.

(Ref: https://robu.in/product/4-channel-isolated-5v-10a-relay-modul)
Figure 3.3: 4-channel relay module

9
3.1.4. Camera Module

The camera module is a critical component in the Smart Plastic Separator system. It is
responsible for capturing high-resolution images of plastic waste as it moves along the
conveyor belt. These images serve as input for the computer vision algorithms, which
analyze and classify the plastics based on their visual features

Features:

• High Resolution: Ensures detailed images for accurate classification.

• Real-Time Capture: Enables the system to process continuous waste streams
efficiently.
• Wide Field of View: Covers the entire conveyor belt to ensure no items are missed.
• Durability: Suitable for industrial environments, resistant to dust and vibrations.

(Ref: https://www.amazon.in/QBM-Camera-Arduino-Sensor-voltage)
Figure 3.4: Camera Module

10
3.1.5. Buzzer Circular Siren LED Warning Light

A Buzzer Circular Siren LED Warning Light is a multifunctional device commonly used
in industrial and safety applications to provide both visual and auditory alerts. It
combines a bright LED warning light with a loud buzzer or siren, making it highly
effective for signaling in noisy environments or critical situations.

(Ref: https://www.amazon.in/Revolving-Industrial-Circular-Warning-LTE-5041-J/)
Figure 3.5: Buzzer Circular Siren LED

3.1.6. DC Power Supply Variable

A Variable DC Power Supply is an electronic device that provides a stable and adjustable
direct current (DC) output. It allows users to set the desired voltage and current levels
within a specific range, making it ideal for testing, prototyping, and powering various
electronic devices..

11
(Ref: 30V 50A VOLTEQ HY3005D Power Supply Coarse and Fine Controls)

Figure 3.7: DC Variable Power supply

12
3.2. Software Requirements

3.2.1. Arduino IDE

The Arduino integrated development environment (IDE) is a cross-platform application

(for Windows, macOS, Linux) that is written in the programming language Java. It is
used to write and upload programs to Arduino compatible boards, but also, with the help
of 3rd party cores, other vendor development boards.

(Ref:https://core-electronics.com.au/tutorials/arduino-ide-tutorial.html)

Figure 3.8: Screenshot of Arduino IDE

13
3.2.2. Visual Studio Code

Visual Studio Code is a source-code editor made by Microsoft for Windows, Linux and
macOS. Features include support for debugging, syntax highlighting, intelligent code
completion, snippets, code refactoring, and embedded Git.

Figure 3.9: Screenshot of VS-code

3.2.3. Google Colab

Google Colab, short for "Google Collaboratory," is a cloud-based platform that allows
users to write and execute Python code directly in their browser. It is particularly popular
among data scientists, machine learning practitioners, and researchers due to its ease of
use and access to free computing resources, including GPUs and TPUs.

14
 Figure 3.10: Screenshot of Google Colab

3.2.4. PyCharm

PyCharm offers out-of-the-box support for Python, databases, Jupyter, Git, Conda,
PyTorch, TensorFlow, Hugging Face, Django, Flask, FastAPI, and more. With the
seamless integration of its context-aware AI Assistant, you can get your work started
quickly and make progress more efficiently. PyCharm equips you with everything you
need for all sorts of projects, from web development and data pipelines to ML
model.prototyping.

15
Figure 3.11: Screenshot of PyCharm

16
CHAPTER 4

METHODOLOGY

4.1 System Overview

The Smart Plastic Separator system is designed to automate the classification and
segregation of plastic waste using computer vision and machine learning technologies. Its
architecture integrates real-time image capture, preprocessing, machine learning-based
classification, and automated sorting mechanisms to efficiently handle diverse plastic
waste streams. High-resolution cameras capture images of plastic waste on a conveyor
belt, which are then preprocessed to reduce noise and enhance features for accurate
analysis. A trained machine learning model identifies and categorizes the plastics,
enabling the automated sorting system to physically separate the waste based on
classification results. Additionally, IoT capabilities allow remote monitoring and data
analysis, ensuring optimal system performance.

4.2 System Flowchart

The system operates in a sequential pipeline, as depicted in the flowchart:

1. Input Stage: Plastic waste is fed onto the conveyor belt.

2. Image Capture: High-resolution cameras capture detailed images of the
waste.
3. Preprocessing: Images are preprocessed to enhance relevant features
(e.g., resizing, noise reduction).
4. Feature Extraction: Computer vision algorithms extract unique features
of the plastics.
5. Classification: The machine learning model classifies plastics into
predefined categories (e.g., PET, HDPE, LDPE).
17
6. Sorting: Servo motors activate the sorting mechanism to separate
plastics based on classification.

Figure 5.1: System flowchart

18
4.3 Dataflow Model

The dataflow model illustrates the movement and processing of data throughout the system:

1. Data Input: High-resolution images captured by the cameras serve as the input.

2. Preprocessing: The raw data undergoes preprocessing to standardize and improve quality.

3. Feature Extraction: Extracted features from images (e.g., color, shape, texture) are passed to

the machine learning model.

4. Classification: Processed data is classified by the trained machine learning model into

specific plastic categories.

5. Output Data: Classification results trigger the sorting mechanism and are logged in the system

database for future analysis and optimization.

19
CHAPTER 6

RESULTS AND ANALYSIS

6.1. Output

The system successfully segregates plastic waste based on visual characteristics, achieving high
accuracy through machine learning algorithms. The computer vision model efficiently classifies
plastics such as PET, HDPE, and LDPE in real time. The integration of the sorting mechanism
ensures the physical separation of classified plastics with minimal errors. The system also provides
real-time logging of operations, aiding in performance analysis and optimization. Demonstrating its
scalability, the system processed diverse plastic waste streams, indicating its potential for industrial
applications.

6.2. Limitations
Despite its success, the system has some limitations:
• Classification Accuracy: Performance may decline when dealing with highly
contaminated or mixed plastic waste.
• Dataset Constraints: The machine learning model's accuracy depends on the size and
diversity of the training dataset.
• Hardware Dependence: The system requires high-performance hardware, which may
limit its adoption in cost-sensitive environments.
• Operational Environment: Factors like inconsistent lighting and camera placement can
impact system accuracy.

6.3. Future scopes

The project opens avenues for future enhancements:

• Expanding the system to classify and segregate a broader range of plastic types.
• Incorporating advanced machine learning models to improve classification speed and
accuracy.
• Integrating IoT for real-time monitoring, remote operation, and predictive maintenance.

20
 • Exploring the use of robotics and automation to further streamline the sorting process.
• Conducting field tests in real-world recycling facilities to evaluate scalability and
performance.

6.4. Work schedule

The project followed a structured timeline to ensure timely completion:

• Phase 1 - Research and Planning: Gathering resources, conducting literature reviews,
and finalizing the project scope.
• Phase 2 - System Development: Designing and implementing the computer vision
model, sorting mechanism, and integration of components.
• Phase 3 - Testing and Optimization: Running test cases to validate system performance,
followed by iterative improvements.
• Phase 4 - Deployment and Analysis: Demonstrating the system's functionality in
simulated environments and analyzing results to refine its effectiveness.

21
CONCLUSION

The "Smart Plastic Separator Using Computer Vision and Machine Learning" project is
an innovative and sustainable approach to addressing the global plastic waste crisis. By
automating the process of plastic waste segregation, the system overcomes the limitations
of traditional methods, which are often labor-intensive, error-prone, and inefficient.
Leveraging advanced technologies such as computer vision and machine learning, the
system efficiently classifies different types of plastics with high accuracy and reliability,
facilitating improved recycling rates. The integration of automated sorting mechanisms
further enhances the efficiency of waste management, significantly reducing reliance on
human intervention. Moreover, the system's real-time processing capabilities and scalable
design make it adaptable for industrial applications, ensuring widespread utility. This
project demonstrates the immense potential of AI-driven technologies in promoting
environmental sustainability by reducing landfill waste, conserving resources, and
mitigating environmental pollution. Despite some challenges, such as hardware
dependency and classification of contaminated waste, the project paves the way for future
enhancements, such as the inclusion of additional plastic types, integration of IoT
features for remote monitoring, and further refinement of machine learning models for
improved performance. By addressing critical environmental issues and contributing to
the advancement of sustainable practices, this project sets a strong foundation for
leveraging artificial intelligence in tackling pressing global challenges, aligning with the
goals of a greener and cleaner future.

22
 REFERENCES

1. Smith, B. Johnson, "Automated Plastic Sorting Using Convolutional Neural

Networks," International Journal of Robotics and Automation, vol. 15, no. 3, 2019.
2. M. Nguyen, D. Lee, "Machine Learning for Plastic Waste Classification,"
Conference on Advances in Sustainable Development, 2020.
3. C. Li, S. Kim, "Enhanced Plastic Detection with Deep Learning Techniques,"
Journal of Environmental Sciences, vol. 34, pp. 102-110, 2021.
4. R. Garcia, L. Wang, "Real-Time Plastic Waste Segmentation Using Computer
Vision," Waste Management Journal, vol. 45, pp. 89-97, 2018.
5. H. Patel, K. Singh, "Intelligent Waste Management System Using Machine
Learning," IEEE Transactions on Industrial Informatics, vol. 17, no. 6, pp. 3819-
3827, 2021.
6. J. Brown, P. White, "Automatic Plastic Bottle Sorting Using Transfer Learning,"
International Conference on Image Processing, 2019.
7. F. Khan, M. Ali, "Deep Neural Networks for Plastic Waste Detection," Journal of
Cleaner Production, vol. 230, pp. 1299-1310, 2020.
8. S. Chen, R. Zhao, "Plastic Waste Classification with YOLO and SSD," International
Journal of Advanced Research in Artificial Intelligence, vol. 9, no. 4, 2021.
9. D. Kim, E. Park, "Integrating IoT and AI for Plastic Waste Management," Sensors
Journal, vol. 18, no. 9, pp. 3265-3275, 2022.

10.A.Gupta, P. Shah, "Optimizing Plastic Waste Sorting with Reinforcement

Learning," Applied Intelligence Journal, vol. 50, pp. 1008-1019, 2021.

23