IoT and AI-Based Early Crop Disease Detection

System with Cloud Dashboard for Farmers

submitted in partially fulfillment of the requirements for the degree of

Bachelor of Technology
in
Electronics and Communication Engineering
by
Vanshika (2200970310185)
Vartika Khare (2200970310186)
Yati Srivastava (2200970310195)

Under the Guidance of

Dr. Monika Bhatnagar
Professor

Department of Electronics and Communication Engineering

([Link] ECE - Accredited by NBA),

Galgotias College of Engineering and Technology, Greater Noida.

(Affiliated to Dr. A.P.J. Abdul Kalam Technical University, Lucknow)

May, 2026

i
 CERTIFICATE

This is to certify that the thesis entitled “IoT and AI- Based Early Crop Disease
Detection System with Cloud Dashboard for Farmers” submitted by Vanshika, Vartika
Khare and Yati Srivastava, Department of Electronics and Communication
Engineering,Galgotias College of Engineering and Technology, Greater Noida affiliated to
Dr. A.P.J. Abdul Kalam Technical University (AKTU), Lucknow, for the award of the
degree of Bachelor of Technology in Electronics and Communication Engineering, is a
record of bonafide work carried out by him under my supervision, as per AKTU code of
academic and research ethics.

The contents of this report have not been submitted and will not be submitted either in part
or in full, for the award of any other degree or diploma in this institute or any other institute
or university. The thesis fulfills the requirements and regulations of the University and in
my opinion meets the necessary standards for submission.

Place : Greater Noida

Date : Signature of the Guide

The thesis is satisfactory / unsatisfactory

I n t e rn a l E xa m i n e r E xt e rn a l E xa m i n e r

Approved by

Head of the Department

ii
 DECLARATION

We hereby declare that the thesis entitled “IoT and AI- Based Early Crop Disease
Detection System with Cloud Dashboard for Farmers” submitted by Vanshika, Vartika
Khare and Yati Srivastava, for the award of the degree of Bachelor of Technology in
Electronics and Communication Engineering to Galgotias College of Engineering and
Technology, Greater Noida affiliated to Dr. A.P.J. Abdul Kalam Technical University,
Lucknow is a record of bonafide work carried out by us under the supervision of Dr.
Monika Bhatnagar.

We further declare that the work reported in this thesis has not been submitted and will not
be submitted, either in part or in full, for the award of any other degree or diploma in this
institute or any other institute or university.

Place : Greater Noida

Date : Signature of the Candidates

iii
 ACKNOWLEDGEMENT

First of all, we would like to express our gratitude to our beloved Chairman Shri. Suneel
Galgotia and CEO Dr. Dhruv Galgotia for proving necessary facilities to carry out and
finish the project successfully. We are grateful to our Director Dr. Brijesh Kumar Singh
for his support and encouragement.

Our special thank goes to Dr. Amrita Rai, Head of the Department, Electronics and
Communication Engineering, for providing an environment that encouraged us in
working towards our goal and supported in completion of our project. This project gave us
an opportunity to apply creative and critical thinking skills.

We would like to present our deep sense of gratitude to our project guide, Dr. Monika
Bhatnagar, Professor, Department of Electronics and Communication Engineering
who has always been a source of motivation and firm support for carrying out the project.
The supervision and support that he gave truly helped in the progression of the project. His
cooperation is much indeed appreciated. We are highly obliged to him for his valuable
advices and moral support during research period.

Finally, our greatest and special gratitude goes to our family for their love and support.

Place: Greater Noida Vanshika

Date: Vartika Khare
Yati Srivastava

iv
 ABSTRACT

Unstable weather conditions and late detection of diseased plants in the fields of India
reduce crop yield and farmer income. One of the main functions of a part is to read live
information (temperature, humidity, moisture of the soil, light) using Arduino UNO
sensors such as DHT11, soil probe, rain detector, and light reader. Unlike sensors working
independently, these information flow into smart models, which monitor fields effectively.
When actual situations do not trigger early alerts, they lead to increased difficult to resolve
issues. In addition, one physical unit on the soil and one logical unit on the web are
presented. The second function of it is an interface that is presented on a web interface with
alerts, graphs and trends, depending on the analysis conducted from data collected.
Artificial Intelligence (AI) with the communication device checks plant parameters rapidly,
increases decision making based on facts and makes sure the result changes before further
losses occur. Therefore, it brings 24 hour supervision on fields with manual scans every
hour replaced. This is nothing like magic; it's just some machines that identify information
human cannot check every few hours. Numbers and statements are placed straight on the
screen; a smart mechanism analyses leaves of plant through pattern analysis using
supervised method (based on known examples), which helps in spotting disease in advance.
The weather information on the webpage stream with real time updating. Decisions made
are based on the fact given by climate data of the fields rather than by an assumption. The
planting schedules will automatically change based on weather variations if weather
conditions change. The farmer has an understandable choice via the webpage, decisions of
watering can be performed in advance and fertilizer usage can be reduced and plant disease
is diagnosed more efficiently than earlier, therefore yield can be increased, waste will be
decreased. The overall yield of the crop is improved because of proper monitoring, and
crop will sustain against unpredictable weather variations.

KEYWORDS: IoT, Artificial Intelligence, Precision Agriculture, Crop Disease Detection,

Convolutional Neural Network (CNN), XGBoost, Cloud Dashboard, Smart Farming,
Arduino, Sensors.

v
 TABLE OF CONTENTS

Chapter CONTENTS Page No.

No.

Acknowledgement iv

Abstract v

Table of Contents vi

List of Figures x

List of Abbreviations xi

1 Introduction 1

1.1 Background and Motivation 1

1.2 Problem Statement 2

1.3 Objectives 3

1.4 Scope 3

1.5 Project Scope 3

1.6 Applications 4

1.7 Limitations 4

2 Literature Review 5

2.1 Introduction 5

2.2 IoT- Based Agricultural Monitoring Systems 5

2.2.1 Role of Sensors in Agriculture 5

2.2.2 Advantages of IoT Systems 6

vi
 2.2.3 Limitations 6

2.3 Artificial Intelligence in Agriculture 6

2.3.1 Machine Learning Techniques 6

2.3.2 Benefits of AI 7

2.4 Crop Disease Detection Techniques 7

2.4.1 Traditional Methods 7

2/4/2 Modern Techniques 7

2.4.3 Convolutional Neural Networks (CNN) 7

2.5 Crop Recommendation System 8

2.5.1 Algorithms applied 8

2.5.2 XGBoost Advantages 8

2.5.3 Limitations 8

2.6 Cloud- Based Agricultural Systems 8

2.7 Integration of IoT and AI 9

2.8 Challenges in Existing Systems 9

2.9 Research Gap 9

2.10 Proposed Approach 9

2.11 Summary 10

3 Methodology 11

3.1 Introduction 11

3.2 System Architecture 11

3.3 Hardware Module 11

vii
 3.3.1 Used components 12

3.3.2 Hardware Working 12

3.3.3 Benefits of Hardware 12

3.4 Software Module 12

3.4.1 Back-End Development 12

3.4.2 Front-End Development 13

3.5 Crop Recommendation Model 13

3.5.1 Algorithm Used 13

3.5.2 Input Parameter 13

3.5.3 Model Working 13

3.5.4 Benefits of the model 13

3.6 Disease Detection Model 14

3.6.1 Algorithm Used 14

3.6.2 Input 14

3.6.3 Model Working 14

3.6.4 Benefits of the model 14

3.7 Data collection and dataset 14

3.8 Data flow and Communication 14

3.9 Integrating IoT and AI 15

3.10 System Working process 15

3.11 Advantages of Proposed Methodology 15

3.12 Limitations of Proposed Methodology 15

3.13 Summary 16

viii
4 Implementation 17

4.1 Introduction 17

4.2 Hardware Implementation 17

4.2.1 Components Used 17

4.2.2 Circuit Design and Connections 17

4.2.3 Working of Hardware System 18

4.2.4 Hardware Testing 18

4.3 Software Implementation 18

4.3.1 Development Environment 18

4.3.2 Backend Implementation 19

4.3.3 Frontend Dashboard 19

4.4 Dataset Collection and Preparation 19

4.4.1 Crop Dataset 19

4.4.2 Disease Dataset 19

4.4.3 Data Preprocessing 19

4.5 Model Training and Implementation 20

4.5.1 Crop Recommendation Model 20

4.5.2 Disease Detection Model (CNN) 20

4.5.3 Model Evaluation 20

4.6 Integration of System Components 20

4.7 Experimental Setup 21

4.8 Performance Analysis 21

4.9 Challenges Faced During Implementation 21

4.10 Summary 21

ix
5 Results and Discussion 22

5.1 Introduction 22

5.2 Hardware Result 22

5.3 Software Result 23

5.4 Crop Recommendation Result 23

5.5 Disease Detection Result 24

5.6 Dashboard outputs 25

5.7 Evaluation 26

5.8 Discussion 26

5.9 Limitations Observed 27

5.10 Summary 27

6 Conclusion and Future Scope 28

References 29- 30

x
 List of Figures

Figure No. Title Page No.

5.1 XGBoost feature importance for crop 22

recommendation model

5.2 Smart Crop Recommendation Data Table 23

5.3 Crop Recommendation Input and Predict Form 23
5.4 Disease Recognition Empty Upload Interface 24
5.5 Disease Recognition Prediction Output Result 24
5.6 Smart Farming Dashboard Homepage View 25
5.7 Dashboard Services and Features Overview 25
5.8 CNN model training vs. validation accuracy and 26
loss over 15 epochs

xi
 List of Abbreviations

AI Artificial Intelligence

API Application Programming Interface

Arduino UNO Arduino UNO R3(Microcontroller Board)

CNN Convolutional Neural Network

CSS Cascading Style Sheets

DHT11 Digital Humidity & Temperature Sensor (Model11)

HTML Hyper Text Markup Language

IoT Internet of Things

LCD Liquid Crystal Display

LDR Light Dependent Resistor

ML Machine Learning

NPK Nitrogen, Phosphorus, Potassium

NumPy Numerical Python

Pandas Python Data Analysis Library

Scikit-learn Scientific Kit for Machine Learning

xii
 CHAPTER-1 INTRODUCTION

1.1 Background and Motivation

Agriculture is the most important sector of the Indian economy and most of the people of
India depends on the business of farming. The traditional form of agriculture is still reliant
on human observations and experience rather than decisions being taken from the data,
leading to poor water utilization, delayed diagnosis of diseases, low yield and so on.

Among the greatest challenges faced by the farmers is late diagnosis of diseases in crops,
where by the time leaves of the crop exhibit noticeable symptoms, a lot of disease has
already spread and due to unpredictable weather conditions, this situation becomes more
aggravated if the diseases are due to insufficient watering.

The application of IOT and AI techniques is a breakthrough for optimizing agricultural

practices. By means of IOT, a real time monitoring of weather conditions and
environmental changes is established by the help of sensors while AI analyzes and
interprets the available data to facilitate data-driven decision making.

The following project covers development of IOT and AI based crop monitoring system.
Data including temperature, humidity, moisture of soil, rain, intensity of light are collected
using sensors that are then processed by an Arduino UNO board and fed into analyzed with
certain tools that present farmers with some beneficial information through AI models.

CNN application is to classify any type of disease at crop life cycle, the disease detected
by the CNN is classified with the help of XGBoost to decide which crop will perform well
with specified conditions. The project is one of the step towards smart farming and to take
the decision based on the past data, future prediction and current data and the Proposed
system consists of Hardware component (IOT) and Software components (AI-Cloud).
There are two Primary Components that Make up the System:

1. Hardware Module IoT System

1
This section of the Hardware Module, which would be at the field location, consist of the:

 Arduino UNO
 Soil Moisture sensors
 DHT11 (Temperature & Humidity sensor)
 Rain Sensor
 LDR (Light Sensor)
 LCD display
 Buzzer & Water Pump

This Hardware module will provide real-time measurements of current environment and
display them onto the LCD display.

2. Software Module AI Based Cloud Dashboard

This section would combine various modules for processing & provide the user with the
Intelligent Insights:

 XGBoost Crop Selection

 CNN Detection of Disease
 API Weather Prediction
 Dashboard Report Visualization

This system will give the Farmer a view into his Crop growth in real-time & act promptly
to the appropriate action required at that time.

1.2 Problem Statement

Although many technology developments have occurred many farmers are not equipped
with a real time agriculture monitoring system. Some of the problems identified are:

 Late detection of crop disease

 Improper irrigation due to no real time soil data

2
  Over-dependence on manual observation
 No single easy-to-use platform integrated monitor
 All of these have leads to loss of crops and wastage of resources.

1.3 Objective

The objectives of the proposed system will be:

 Monitor soil and environmental parameters by IoT sensors

 Build an integrated real-time cloud dashboard
 Use data to help farmers make more effective decisions.

1.4 Scope

The area of research of this project will consist of developing & testing a smart agriculture
system by using a sample dataset and simulated data. The system is capable of real-time
dashboard visualization, detecting crop diseases, weather forecasting, & recommending
the proper crop. Future enhancements may consist of implementing real time hardware and
adding the capability to use a mobile application.

1.5 Project Scope

The aim of this project is to give the users a smart, integrated approach to farming by
utilizing both the capabilities of IoT and AI to ensure the optimization of agriculture
practices. The project will encompass:

 Real-time monitoring of the state of a farm.

 AI based image detection of disease within crops.
 Recommendations of suitable crops to grow based on environmental conditions.
 A web-based cloud dashboard for viewing of system data.
 However, the current implementation of the project will only consist of the
following:
 Prototype implementations

3
  Use of pre-existing datasets
 Limited sensors range

The future enhancements of the project can involve the utilization of real time components
and more advanced AI models to create a more efficient system.

1.6 Applications

The application can be widely used in many areas of life, including:

 Smart farming systems

 Precision agriculture
 Crop monitoring solutions
 Agriculture research
 Automated irrigation systems

The farm system will allow farmers to take data-based decision to produce more.

1.7 Limitations

Despite some positive aspects, the system also consists of certain limitations that should
be taken into account:

 The output will depend on the quality of the dataset.

 The cloud dashboard would necessitate the internet to be accessible at the field.
 Under harsh weather conditions (for example extreme freezing), sensors might
report incorrect data.
 It may misclassify similar diseases when using a CNN model.
 Initial setup of the system may require a large upfront investment.
 It may be possible to overcome these limitations in future system

4
 CHAPTER 2 LITERATURE REVIEW

2.1 Introduction
Agriculture has been the fundamental component in the economic development of all the
countries. However, several traditional farming techniques face numerous issues such as
altering climate, barren land, a lack of water, the spread of plant diseases, and so forth.
These problems directly affect the main productivity of agriculture (e.g. The output level
of the crop and the incomes of farmers).
Digital technologies and automation have been used for decades to advance the agricultural
sector. Today it is evolving toward smart farming that makes decisions based on real-time
data instead of gut feelings. Technologies such as IoT, AI, ML, and Cloud Computing play
a critical role in this progression [1, 3].
Chapter 2 would present the literature reviews of the following research works in
agricultural management; namely IoT crop monitoring, AI-based crop analysis, techniques
of disease detection in crop, recommending the appropriate crop and cloud based
agricultural platform. Some of the limitations will be highlighted, along with the research
gap suffered by this project.

2.2 IoT-Based Agricultural Monitoring Systems

IoT technology is highly effective for collecting and using data in agriculture. It is based
on network of sensors and electronic devices monitoring environment condition in real
time [1, 3].

2.2.1 Role of Sensors in Agriculture

Sensors play a significant role in IoT systems. Typical sensors in use include:
 Temperature sensor
 Humidity sensor
 Soil moisture sensor
 Rain sensor
 Light intensity sensor

5
These sensors would be positioned out in the fields and interfaced with microcontrollers
like Arduino and Raspberry Pi.

2.2.2 Advantages of IoT Systems

 Continuous monitoring of environmental condition
 Reduction in manpower
 Improvement in irrigation management
 Better usage of resources

2.2.3 Limitations
However, some limitations to IoT systems include:
 No intelligence in decision making
 Dependence on the internet connection
 High initial costs of implementation
 Accuracy issues under severe environmental conditions
 The current majority of IoT platforms do not incorporate forecasting [1].

2.3 Artificial Intelligence in Agriculture

Artificial Intelligence enables machines to analyze data and make human-like decisions.
These include [3, 4]:
 Crop prediction
 Detection of diseases
 Forecasting of yield
 Analysis of soil condition

2.3.1 Machine Learning Algorithms

Machine learning algorithms work on past information to predict future events by pattern
recognition. They are divided into;
 Supervised learning
 Unsupervised learning

6
  Reinforcement learning

2.3.2 Benefits of AI
 Improved accuracy of forecasts
 Decreased manual work
 Speed of decisions made
 Effective management of resources
 As observed, AI system requires high quantities of data and training [4].

2.4 Crop Disease Detection Techniques

The image above shows a canker occurring on broccoli, where this disease which causes it
would not typically be clearly distinguishable from other similar diseases. The bacteria
cyanobacteria can also result in this.

2.4.1 Traditional Methods

Traditional methods are:
 time consuming
 not accurate
 not scalable

2.4.2 Modern Techniques

Modern techniques utilize image processing techniques along with deep learning.

2.4.3 Convolutional neural networks (CNN)

One very commonly used technique on image classification is the CNN or convolutional
neural network that would be capable to extracting relevant features from images to classify
them [2, 5, 8, 9, 12, 13].
Advantages
 High accuracy
 Automatic feature extraction

7
  Early detection

Limitations
 Large amount of data is needed
 Sensitivity to variations in illumination
 High computational demand

2.5 Crop Recommendation Systems

Crop recommender system suggests optimal crop choice given a set of inputs [6, 7, 10, 11,
14, 15].

2.5.1 Algorithms applied

 Decision Tree
 Random Forest
 Support Vector Machine
 XGBoost

2.5.2 XGBoost Advantages [6, 7, 15]

 High accuracy
 Speedy computation
 Good management of large data sets

2.5.3 Limitations
 Requires effective feature selection
 Effectiveness dependent on the data set

2.6 Cloud-Based Agricultural Systems

The cloud can be leveraged as a supporting layer in this project and it can provide
processing and storage needs for Big Data [1, 3].
Features

8
  Remote data access
 Real-time analytical capabilities
 Scalability

Applications
 Data storage
 Data visualization on dashboards
 Decision support systems

2.7 Integration of IoT and AI

When combined, these two technologies create a powerful intelligent system that can
perform the following tasks:
 Monitoring
 Prediction
 Automation of decisions
While integration still comes with challenges, along with the costs associated with it [1, 3].

2.8 Challenges in Existing Systems

 No real-time integration
 High implementation costs
 Lack of training data
 Environmental variability issues
 Sensor reliability

2.9 Research Gap

 Observations of previous works in the field are that:
 The system's design is often based on IoT, but not on AI, or on AI and not on IoT.
 There are not many real-time decision-making systems.
 Lack of user friendly dashboards.

9
2.10 Proposed Approach
The developed system will utilize:
 IoT sensors for data monitoring.
 AI models for predictions and forecasting.
 Cloud dashboard for displaying results and visualization.

2.11 Summary
In this chapter, the various current smart agriculture technologies have been evaluated and
the need for a combined system has been established.

10
 CHAPTER 3 METHODOLOGY

3.1 Introduction
In this chapter we have proposed and design an IOT and AI based system for Early
Detection of Crop Diseases that monitor the environmental parameters with the help of
various sensor and recommend some strategies to the farmer using AI technologies. The
system composed of the combined system of IoT and AI for creating a smart agriculture
[Link] work involved the data transmission from devices to the system and a
prediction [Link] flow is:
Data Collection->Data Transmission->Data Processing->Model Prediction->Result
display

3.2 System Architecture

The proposed system comprised of three layers, listed below:
1. Data acquisition layer
All sensors and hardware that fetches all real-time environmental data from the field would
be part of this layer.

2. Processing Layer
Hardware device(Arduino) and cloud servers, which are used for processing all acquired
data as well as the prediction models will comprise of this layer.

3. Application Layer
The user would see sensor data and all predictions on this web based dashboard.
Overall flow:
Sensor Data --> Arduino --> Processing --> Cloud server --> AI models --> Dashboard --
> Output.

3.3 Hardware Module

This module would used to collect the data for all environmental factors.

11
3.3.1 Used Component
Arduino UNO- This is the control center of our system.
DHT11 Sensor- This sensor is used to measure temperature & humidity.
Soil Moisture sensor- This is used to determine the amount of water in the soil.
Rain Sensor- this is used to detect rain.
LDR (Light dependent resistor)- This is used to measure the light received by it.
LCD Display- Used to display sensor data on real time.
Buzzer- Used to notify if conditions are not ideal.
Water Pump- An optional component used for irrigation automation.

3.3.2 Hardware Working

 Sensor would be receiving continuous inputs from the environment.
 Arduino receives input from sensors using its digital/analog ports.
 The received input from Arduino would then be displayed on the LCD screen.
 If the conditions are not ideal (i.e. Low soil moisture) the buzzer would be activated.
 All the input collected would then be transmitted to the cloud server for the models
to utilize.

3.3.3 Benefits of Hardware

 Real-time monitoring
 Automated alarm
 No human monitoring is needed.

3.4 Software Module

This software module is used to make recommendations and detection of the disease from
the collected data.

3.4.1 Back-End Development

This module has been developed using Python and Flask, that is used to all communications
between hardware and AI models and also used for data processing.

12
3.4.2 Front-End Dashboard
The web-based dashboard would involve the display of;
 Sensors data
 Recommendations on crops
 Results for disease detection
 Weather forecast would all be displayed here.

3.5 Crop recommendation Model

The model would be used for suggesting crops according to the present conditions of
weather and soil.

3.5.1 Algorithm Used

XGBoost algorithm is used because it is the highly used algorithm due to the fast
performance and accurate results.

3.5.2 Input Parameter

 Temperature
 Humidity
 Soil moisture
 Rainfall

3.5.3 Model working

The collected sensor data are fed into the XGBoost algorithm that in return recommends
the most efficient crop.

3.5.4 Benefits of the model

 Accurate
 Fast predictions
 Handles a large dataset.

13
3.6 Disease Detection Model
The model is designed to help diagnose the disease in a crop.

3.6.1 Algorithm used

CNN algorithm is used because it is excellent in image analysis, it is used here for detection
and classification of plant disease using image.

3.6.2 Input
Input to the model are images taken from the leaves of the crop.

3.6.3 Model Working

The images taken would be preprocessed, then fed to the CNN algorithm which detects the
disease.

3.6.4 Benefits of the model

 Early detection of disease
 Highly accurate
 Automatic feature extraction.

3.7 Data collection and dataset

It is data sets which the machine learning models are trained on.
Data Sets used:
 Crop data (taken from Kaggle)
 Image dataset for plant diseases

Data Preprocessing:
 Data Cleaning.
 Scaling / Normalization.
 Feature selection.

14
3.8 Data flow and Communication
The data flow would occur in the following sequence:
1. Sensors gather real-time data.
2. Data are transmitted to the Arduino.
3. Arduino transfers the data to the cloud server.
4. AI models interpret and use this data.
5. Recommendations are displayed on the dashboard.

3.9 Integrating IoT and AI

Combination of the IoT and AI devices:
1. IoT devices would be the one collecting all data from the field.
2. AI models would utilize these inputs and produce an intelligent recommendation.
3. Recommendations would be displayed to the user on the dashboard.

3.10 System Working process

The system would be working in following steps:
1. Sensors would continually monitor weather conditions.
2. Sensor data would be transferred to the Arduino.
3. Arduinos will send data to cloud server
[Link] model will learn input data.
5. The results of analysis would be displayed on the dashboard.
6. Farmers take action on the basis of displayed recommendations.

3.11 Advantages of Proposed Methodology

 Real time monitoring
 Early diagnosis of diseases
 Increased yield
 Resourceful use of resources
 User-friendly interface

15
3.12 Limitations of Proposed Methodology
 Requires Internet connectivity
 Sensor errors in reporting values
 Accuracy is dependent on dataset.
 Cost for setting up.

3.13 Summary
In this chapter we have discussed the entire methodology required for developing the smart
agriculture system. We talked about hardware components, software modules, and the AI
models along with the data flow for communication between different components. Both
the AI and IoT devices contribute to building an efficient and reliable system.

16
 CHAPTER 4 IMPLEMENTATION

4.1 Introduction
In this chapter we will introduce the proposed implementation of IoT and AI-based Crop
Disease Detection System in practice. We will examine both the hardware and software
aspects of the system and outline the procedure for preparing the dataset, training the
model, and configuring the experiment.
The core idea behind implementing this system is to combine IoT devices and AI models
to create a system that can monitor the crop field in real-time and make intelligent
decisions. The experimental setup of the system was designed to assess the performance of
both the hardware and the software in a real-world environment.

4.2 Hardware Implementation

The hardware is an important part of the system because this component gathers the data
from the field.
4.2.1 Components Used
The following components were used for the hardware implementation:
 Arduino UNO: A microcontroller that processes sensor data.
 DHT11 sensor- To record the temperature and humidity of the area.
 Soil Moisture sensor- To read the moisture level of the soil.
 Rain sensor- To detect the presence of rain.
 LDR (Light Dependent Resistor): Measures light intensity.
 LCD Display: Shows the current sensor readings.
 Buzzer: Alerts users to any abnormal conditions.
 Water Pump (optional): Automates irrigation.

4.2.2 Circuit Design and Connections

Sensors are plugged into the Arduino using both analog and digital pin.
The system is powered by the Arduino's USB port and an external power adapter.

17
The LCD is connected to the Arduino using an appropriate interface (parallel or I2C).
The buzzer is connected to one of the Arduino's digital pins.
The circuit diagram is designed to provide accurate and reliable sensor readings while
minimizing interference and noise.

4.2.3 Working of Hardware System

1. Sensors continuously monitor the environmental parameters of the field.
2. The Arduino reads sensor data at specific intervals.
3. Sensor readings are shown on the LCD display
4. The buzzer is activated when the soil moisture level is less than the specified threshold
5. All sensor data is transmitted to a cloud server.

4.2.4 Hardware Testing

The hardware was tested under various conditions, and the sensor readings were compared
to expected values. The results were found to be accurate.

4.3 Software Implementation

The purpose of the software section is to handle the data and displays the results.

4.3.1 Development Environment

 The software system was developed with the following tools:
 Programming Language: Python
 IDE: Jupyter Notebook / VS Code
 Framework: Flask
 Libraries:
 NumPy
 Pandas
 Scikit-learn Tensor Flow / Keras

4.3.2 Backend Implementation

18
The backend system has the following functionalities:
 Receives sensor data.
 Processes the sensor data.
 Executes the AI models.
 Sends the results to the front-end.
The Flask framework was used to create APIs to integrate the hardware system with the
front-end.

4.3.3 Frontend Dashboard

The output of the system is presented to the user via a web based interface. It consists of:
 A live display of data from the sensors.
 Crop recommendations.
 Crop disease detection results.
 Weather information.
The aim of the dashboard was to create a user friendly environment for the farmer.

4.4 Dataset Collection and Preparation

The quality and the amount of data trained in the AI model is also what makes the model
to perform well.

4.4.1 Crop Dataset

The crop dataset contains different climatic values like temperature, humidity, and rain and
recommended crop for that condition. This data is taken from Kaggle.

4.4.2 Disease Dataset

The disease data set consists of images of healthy and diseased plant leaves, and disease
class for each disease. This was the training data set for the Convolutional Neural Network.

4.4.3 Data Preprocessing

The data preprocessing steps included:

19
  Data Cleaning.
 Data Normalization.
 This divides the data into a training set and a test set.
 Resizing and normalizing image data.

4.5 Model Training and Implementation

4.5.1 Crop Recommendation Model (XGBoost)

Split dataset into training and testing set
 Trained the XGBoost model on the training set
 Hyper parameters were tuned to get a better performance for the model

4.5.2 Disease Detection Model (CNN)

 Images were preprocessed.
 The image dataset with labels was used to train the CNN.
 The model was trained on a varying number of epochs in order to obtain high
accuracy.

4.5.3 Model Evaluation

The performance of the models was evaluated using Accuracy, Precision, Recall and F1-
score.

4.6 Integration of System Components

The system includes hardware, back-end, front-end. All the components are integrated
together as one system.
1. Hardware receives the data from the field.
2. All received data is transferred to the cloud server.
3. The data is processed using the AI models.
4. The results are displayed on the front-end.

20
4.7 Experimental Setup
The experimental setup included:
 Deployment of the hardware prototype in the field.
 Testing of the sensors under various environmental conditions.
 Testing of the models on the test data set.

Test conditions included:

 Wet and dry soil moisture levels.
 Different temperature and humidity levels.
 Various light conditions.

4.8 Performance Analysis

The evaluation criteria for the system:
 Sensor accuracy.
 Disease prediction accuracy of the models.
 System response time.
 The system performed well under the test conditions.

4.9 Challenges Faced During Implementation

 Sensor calibration might be an issue.
 The inconsistency in collected data poses a challenge.
 There is a risk of overfitting the model during training.
 The complexity of integrating all the components is also a challenge.
These challenges were addressed through extensive testing and model training.

4.10 Summary
The chapter discusses the entire implementation of the proposed system: Hardware
development, Software development, and AI model training and testing. IoT and AI
combine to provide a more intelligent agriculture solution

21
 CHAPTER 5 RESULTS AND DISCUSSION

5.1 Introduction

The obtained values were tested the IoT and AI based Crop Disease Detection system is
discussed in this chapter with a combination of hardware and software performance
analysis along with the plots. The results obtained is discussed on the basis of the
performance of: sensors, model, and system. The effectiveness in making decision in
farming field is discussed here.

5.2 Hardware result

Various hardware configurations are tested for various ambient changes and test results for
sensor testing performance are displayed.

 The sensor outputs for temperature and humidity were quite stable even for
different environment readings.
 The soil moisture sensor response correctly for wet and dry situation.
 The rain sensor is able to detected presence of water fine.
 The LDR was capable of detecting light change very fine.
 When the soil moisture drops to a value below the threshold, the Buzzer could buzz.

Fig 5.1-XGBoost feature importance for crop recommendation model

22
5.3 Software Result:
There are two parts within software module, they are

 Crop Recommendation model

 Disease Detection model

5.4 Crop Recommendation Result:

XGBoost model was trained using environmental data and testing on its prediction

accuracy.

Fig 5.2 – Smart Crop Recommendation Data Table

Fig 5.3 – Crop Recommendation Input and Predict Form

23
5.5 Disease Detection Results
For the detection of disease the module will take CNN model which would classify an
image taken from a plant leaf into diseased or healthy. While training it was already
categorized and the model turned out to be very effective in identifying the disease. While
we see the graphs showing increase in accuracy and decreasing loss, it is clear that the
CNN model was indeed learning, which is a good learning measure. We would receive
good value for performance metrics: precision, recall and F1-score. It can identify the
disease in early stage to assist the farmer to fix the disease at the earliest. Thus it is an
effective system for detection of crop disease.

Fig 5.4 – Disease Recognition Empty Upload Interface

Fig 5.5 – Disease Recognition Prediction Output

24
5.6 Dashboard outputs
All the above outputs are combined together and represented to the user through a user-
friendly graphical user interface(dashboard).

Observable properties:

1. Real-time sensor data

2. Crop recommendations
3. Disease detection results
4. Weather data

Fig 5.6 – Smart Farming Dashboard Homepage View

Fig 5.7 – Dashboard Services and Features Overview

25
5.7 Evaluation

System performance is measured on some common metrics:

 Accuracy: Number of correctly made predictions.

 Precision: Proportion of correct positive predictions that actually have the class that
it's predicted for.
 Recall: Proportion of actual positives that were correctly identified.
 F1-score: Harmonic mean of precision and recall.

Fig 5.8-CNN model training vs. validation accuracy and loss over 15 epochs

Observation:
Both models very accurate.
Sensors data seems accurate enough.
Very quick to react.

5.8 Discussion
From these results we can observe that the designed system can be executed successfully
in a real/situational environment. Important achievements in application:
 Disease early detection

26
  Efficient water utilization
 Optimal crop selection
 Less manual work

Implications for farmers:

 Increased productivity
 Minimizing losses
 Efficient decision making

The application of IoT and AI technology, makes the proposed system smart, feasible.

5.9 Limitations Observed

 Sensor errors in adverse environments

 Requirement of internet connection
 Relatively small dataset size
 CNN might not be able to identify similar diseases correctly

5.10 Summary

In this chapter we discussed the results of proposed system and examined its accuracy.
The figures as given above demonstrates the efficient, accurate and reliable nature of the
system.
This chapter concludes the project work and explains future direction of it.

27
 CHAPTER 6 CONCLUSION AND FUTURE SCOPE

Conclusion

It can be concluded from the project that, a real time monitoring system using IoT and AI
has been designed to monitor crop condition which provided benefits over the old style
agriculture system. The system was found to detect and predict the crop and resources
needs in order to have better harvest with integration of machine learning for decision
support system and deep learning for diseases prediction at early stage. This smart crop
monitoring system is user friendly and it can add a lot to smart agriculture.

Future Scope:

To increase the system's precision and assist the farmer in making appropriate decisions,
high-level sensors for sensing soil nutritional information and soil pH could be included.
In addition to this, instant notifications and direct control of automatic irrigation system
can be provided through the mobile application. To reduce cost incurred by users, the
system can make use of solar energy. The AI model could be trained with higher efficiency
by including more data sets, and be further developed into different language so as to
include larger number of users.

28
 REFERENCES

[1]. Kumar, V.; Sharma, K. V.; Kedam, N.; Patel, A.; Kate, T. R. (2024). A comprehensive
review of smart and sustainable agriculture using IoT. Computers and Electronics in
Agriculture, 210, 107895 [6].

[2]. Kumar, D., & Rajesh, S. (2024). Deep neural networks for multi-crop disease detection
using IoT cameras.

[3]. Ahmed, N. (2024). Advancing agriculture through IoT, Big Data, and AI. Computers
and Electronics in Agriculture, 215, 107973.

[4]. A. Kar, N. Nath, U. Kemprai, and Aman, "Performance Analysis of Support Vector
Machine (SVM) on Challenging Datasets for Forest Fire Detection," International Journal
of Communications, Network and System Sciences, vol. 17, no. 2, pp. 11-29, Feb. 2024.

[5]. Balamurugan, K., & Raj, S. (2023). Deep learning-driven plant health monitoring using
IoT and vision sensors.

[6]. K. Jhajharia, P. Mathur, S. Jain, and S. Nijhawan, "Crop Yield Prediction using
Machine Learning and Deep Learning Techniques," Procedia Computer Science, vol. 218,
pp. 406-417, 2023.

[7]. C. Raju, A. D, and A. P. B, "CropCast: Harvesting the future with interfused

MachineLearning and advanced stacking ensemble for precise crop prediction," Kuwait
Journal ofScience, p. 100160, Dec. 2023.

[8]. Md. Manowarul Islam et al., "DeepCrop: Deep learning-based crop disease prediction
with web application," Journal of Agriculture and Food Research, vol 14, pp. 100764-
100764, Dec. 2023.

29
[9]. D. Tirkey, K. K. Singh, and S. Tripathi, "Performance analysis of AI- based solutions
for cropdisease identification, detection, and classification," Smart Agricultural
Technology, vol. 5, p. 100238, Oct. 2023.

[10]. M. Shripathi Rao, A. Singh, N. V. Subba Reddy, and D. U. Acharya, "Crop prediction
using machine learning," Journal of Physics:

[11]. M. Shripathi Rao, A. Singh, N. V. Subba Reddy, and D. U. Acharya, "Crop prediction
using machine learning," Journal of Physics: Conference Series, vol. 2161, no. 1, p.
012033, Jan. 2022.

[12]. G. Rani, E. T. Venkatesh, K. Balaji, Balasaraswathi Yugandher, Adiki Nithin Kumar,

and M Sakthimohan, "An automated prediction of crop and fertilizer disease using
ConvolutionalNeural Networks (CNN)," 2022 2nd International Conference on Advanced
Computing and Innovative Technologies in Engineering (ICACITE), Apr.2022.

[13]. A. Bouguettaya, H. Zarzour, A. Kechida, and A. M. Taberkit, "A survey on deep

learning-based identification of plant and crop diseases from UAV-based aerial images,"
Cluster Computing, Aug. 2022.

[14]. K. Ramu and K. Priyadarsini, "A Review on Crop Yield prediction Using Machine
Learning Methods," 2021 2nd International Conference on Smart Electronics and
Communication(ICOSEC), Oct. 2021.

[15]. S. M. PANDE, P. K. RAMESH, A. ANMOL, B. R. AISHWARYA, [Link],

and K. SHAURYA, "Crop Recommender System Using Machine Learning Approach,"
2021 5th International Conference on Computing Methodologies and Communication
(ICCMC), Apr. 2021.

30
IoT and AI-Based Early Crop Disease
Detection System with Cloud Dashboard
for Farmers
Vanshika, Vartika Khare, Yati Srivastava & Dr. Monika Bhatnagar

Department of ECE, Galgotias College of Engineering and Technology, Greater Noida, Uttar
Pradesh, India, 201310

ABSTRACT: The agriculture is important and critical sector in India, but unfavorable weather conditions
continue to damage the crop yields, since the crop diseases tend to go unnoticed until the damage is too
severe- this lowers yield and puts farmers into financial strain. A novel research addresses the above problem
through integrated use of the advanced technologies, such as IoT and AI, thus to enable the real-time updates
for decisions to be made rather than estimates. This research makes use of two parts, which can either monitor
the crops using a hardware part, or the resulting information can be visualized on a software part. In the
fields, a hardware part focused around an Arduino UNO collects continuously collected data; ambient and
soil moisture levels, temperature, rain indicator, and light readings, which can be observed directly on a
miniature screen situated within vicinity. Images of crops are passed on to a high resolution image processing
network which detects the diseases, often before they are visible to naked eye. In parallel, an XGBoost
algorithm utilizes the weather conditions to suggest the type of seed to be used under present conditions.
Farmers can obtain useful, easily digestible data on the screen as it is rendered on a web framework and
hosted online. Real-time weather data can also be obtained for better irrigation decisions. Feed suggestions
can be more precise. Disease control can be implemented through visual observation and detection, rather
than estimation. Crop yields can be maximized, and resources can be saved.

KEYWORDS: IoT, Artificial Intelligence, Precision Agriculture, Crop Disease Detection, Convolutional
Neural Network (CNN), XGBoost, Cloud Dashboard, Smart Farming, Arduino, Sensors.

I. INTRODUCTION
In India, Farming makes up nearly all lives and thousands of people move throughout the country villages
and farms. Old ways, or decisions based on hunches rather than figures slow progress down, and when
estimates are used instead of calculations then there's no catching problems until they have gotten so bad,
that even changing the rainfall in an area, doesn't help, and the right irrigation is no longer known, the biggest
problem comes in when disease strikes in a crop, as often that cannot be seen until there is yellowing in the
leaves or spotted in the plants themselves, at this stage much damage has been done, and the roots of the
plants are beginning to weaken, harvests are declining as the grower sprays vast amounts of chemicals, but
the damage is usually irrevocable from this point onward, leading to decline in produce and quiet losses, [1]
There out here where soil meets figures, a new methodology of farming quietly infiltrates into crop
cultivation. In place of gut feelings about what irrigating practices should be carried out, there are in place,
in situ measurements taken minute by minute. There are even constant measurements of changes in
temperatures, or moisture conditions, which can then be observed in patterns from which conclusions may
be drawn using only an intelligent piece of software. Image sensors can then combine to the mixture to feed
the computer with an image of the soil, and over a period of time the machine learns. It comes one piece of
evidence at a time though there is no such thing as speed, rather a careful analysis is carried out.

31
Acceptance Mail

32
 Page 2 of 46 - Integrity Overview Submission ID trn:oid:::3618:137200504

16% Overall Similarity

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

Filtered from the Report

Bibliography

Match Groups Top Sources

51 Not Cited or Quoted 16% 12% Internet sources

Matches with neither in-text citation nor quotation marks
9% Publications
2 Missing Quotations 0% 14% Submitted works (Student Papers)
Matches that are still very similar to source material

2 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
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 46 - Integrity Overview Submission ID trn:oid:::3618:137200504

 VANSHIKA
+917761053136 | vanshikanandani2@[Link] | [Link]/in/vanshika-nandani |
[Link]/vanshika-nandani

Career Objective
With a strong foundation in electronics and communication systems, along with hands-on experience in web
development. Passionate about learning new technologies and working with teams to deliver impactful web
solutions, combining both hardware and software to create innovative results.

Educational Qualifications
• Pursuing Bachelor of Technology in Electronics and Communication Engineering at Galgotias
College of Engineering & Technology, Greater Noida, with a CGPA of 7.34 (up to 6th semester),
graduating in 2026.
• Pursuing Data Science Program (Online) from IIT Madras.
• Completed Class XII from S.G.D.M College, Patna, affiliated with BSEB, securing 78% in 2021.
• Completed Class X from Dr. G.L. Dutta D.A.V Public School, Patna, affiliated with CBSE, securing 77.8%
in 2019.

Internship
Web Development Intern at Zidio Development [May 2024 – August 2024]
Roles and Responsibilities: Worked on the frontend development of a web application using HTML, CSS,
and JavaScript. Designed and implemented responsive and user-friendly interfaces to enhance the user
experience. Collaborated with the development team to integrate features and updates, and managed version
control by pushing code to GitHub.

Projects
KrushtiTech – IoT and AI-Based Smart Agriculture System [Ongoing]
• Developed AI-based modules for crop recommendation and plant disease detection using Python,
TensorFlow, and OpenCV.
• Designed a Streamlit web dashboard for farmers to access crop insights and disease analysis.
• Customized the frontend interface using HTML, CSS, and JavaScript for a user-friendly experience.
• Designed the system architecture for future integration with IoT sensors (ESP32) to enable real-time soil
monitoring.

Rain Sensing Hand Free Shed Using Arduino Uno [October 2023 – January 2024]
• Designed and developed a working model that automatically detects rainfall and operates the shed roof
without manual control.
• Utilized Arduino UNO, rain sensor, and DC motor for automated movement and real-time response.
• Enhanced convenience and protection for stored items through a hands-free control mechanism based on
sensor feedback.

Skills
Technical Skills: Python, HTML / CSS, JavaScript, AI&ML, Gen AI, VS Code.
Soft Skills: Problem-solving, Teamwork, Communication, Adaptability, Time Management, Critical Thinking
 VARTIKA KHARE
LinkedIn: [Link]/in/vartika-khare Email: vartikakhare732003@[Link]
Mobile: (+91) 8604789326

SUMMARY

A passionate final-year ECE student with a strong academic foundation in digital communication, signal processing, and wireless networks (4G/5G),
with hands-on experience in Drone systems and Arduino, eager to apply technical expertise in electronics and programming to real-world
engineering challenges, while continuously learning and contributing to innovative solutions.

EDUCATION
Electronics and Communication Engineering | Galgotias College of Engineering and Technology Current GPA- 7.59 | June 2026
XII (CBSE) | Shri Ram Public School 92.4% | June 2021
X (CBSE) | Oxford Model Senior Secondary School 91% | June 2019

SKILLS
Programming Languages: Python
Core Subjects: VLSI, DC, DSD
Database System: MS Excel, MySQL
MS Office: Word, Excel, PowerPoint
Drone Technology: Flight controllers, Propulsion systems
RC Transmitter / Receiver Systems: Futaba, FlySky
Tools: Visual Studio Code, MATLAB, Arduino IDE, Falcon Design Space, Google Teachable, Scratch Jr and 3.0
CAD Software: TinkerCAD, Creality Print
Soft Skills: Problem Solving, Team Work, Project Management, Technical Communication

PROJECTS
An IoT and AI-based early crop disease detection system with a cloud dashboard for farmers Nov 25- March 26
 Real-time and accurate monitoring of soil and environmental parameters.
 Efficient utilization of water and fertilizer resources.
 Early detection and prevention of crop diseases.
 User-friendly interface for monitoring and decision-making.
 Enhanced crop productivity and sustainable farming practices.
.
Rain Sensing Hand Free Shed Dec 23- Jan 2024
 An Arduino Rain Shelter efficiently deploys a protective shed upon rain detection, offering a user-friendly and automated solution for
staying dry during unexpected rains. It automatically opens and closes the shelter based on weather conditions, offering convenience and
protection from unexpected showers.
 Key benefits include automatic protection, enhanced safety, convenience, and efficiency.
 Individuals often face the inconvenience of being caught in the rain while working in an outdoor environment, emphasizing the necessity
of a hands-free rain-sensing shelter.

POSITION OF RESPONSIBILITY
Placement Coordinator | CPDD April 2025- Present
Placement Coordinator | T&P April 2025- Present
Head | Literacy Club of GNIX Society Sept 2024- May 2025
Head | Dramatic Society of GCET Nov 2024- May 2025
Coordinator | Dramatic Society of GCET Oct 2023- Aug 2024
Coordinator | Marketing and Promotions Club Sept 2023- Aug 2024
Member | Dramatic Society of GCET Feb 2023- Aug 2023
Member | Marketing and Promotions Club Feb 2023- Aug 2023

CERTIFICATES

Google Cloud - Generative AI | Smart Bridge August 2025

GoFr Summer of Code August 2025
Generative AI Badge | Google Cloud July 2025
Web Development Fundamentals | IBM July 2025
HTML & CSS Boot camp | Lets Upgrade July 2025
Network Bulls Certificate April 2025
Sustainability E Learning | Nestle March 2024
Cybersecurity Essentials | CISCO June 2023
Introduction to Cybersecurity | CISCO June 2023
 YATI SRIVASTAVA
Uttar Pradesh,India
yatiprem691@[Link] | +919565550988|LINKEDIN
EDUCATION
GalgotiaCollege of Engineering and Technology 2022 - 2026
Noida,India (expected)
Bachelor of Technology - Electronics & Communication
Courses: Digital Signal Processing, Embedded Systems, VLSI Design, Control Systems, Communication Systems, Microprocessors &
Microcontrollers, Electromagnetic Field Theory.
RAM SEWAK RAM KUMAR I.C. , SiddharthNagar
2021-2022
UttarPradesh,India
Percentage- 70.4% | Matriculation (Class XII)

St. Basil School

2018 - 2019
UttarPradesh,India
Percentage- 60.83% | Intermediate(Class X)

SKILLS
Programming & Web Development: Java, C++, HTML, CSS, JavaScript, OOPS (java),
Version Control & Tools: GitHub, Microsoft Excel (Basic), Microsoft Word, Microsoft PowerPoint, Google Workspace
Design & Presentation: Canva (Graphic Designing), UI/UX Basics, Presentation Design, Creative Thinking
Soft Skills & Professional Traits: Leadership, Critical Thinking, Team Collaboration, Problem Solving, Time Management, Adaptability
Languages: English (Fluent), Hindi (Fluent)

EXPERIENCE
Junʼ24 - Sepʼ24
DiloxyTechnology Private Limited
Web Development Intern
Worked on creating clean, responsive web pages using HTML and CSS, with a focus on mobile-friendly layouts and compatibility across different browsers.
Collaborated with the design team to build engaging templates and visual content using Canva, contributing to client presentations and social media
campaigns.
Gained practical experience in developing interactive components with Java and explored the use of WordPress for building and customizing websites.
Took part in regular code reviews and team discussions, which helped me understand real-world development workflows and UI/UX best practices.
Helped maintain and update the company’s websites, ensuring smooth user experience and consistent branding throughout.
Nexteir Technologies Private Limited
Full Stack Developer Intern
Worked as a Full Stack Developer Intern for 3 months, contributing to the design, development, and deployment of live web-based projects.
Actively participated in the development of a logistics platform, implementing front-end and back-end functionalities to improve system performance
and user experience.
Gained hands-on experience with modern technology tools, development methodologies, and software development processes.
Collaborated with cross-functional teams to analyze requirements, develop features, debug issues, and optimize application workflows.
Demonstrated strong problem-solving skills, technical proficiency, teamwork, and commitment to project delivery in a professional work environment.

PROJECTS
RainSensing Free Hand Umbrella
Automatic Activation: The umbrella uses rain sensors to detect precipitation and automatically opens when it starts raining, offering hands-
free protection without manual operation.
Smart Design for Convenience: Designed for users with occupied hands (e.g., carrying bags or using a phone), it typically incorporates
wearable or backpack-mounted components to keep the user dry without the need to hold the umbrella.

ACADEMIC AND EXTRACURRICULAR ACHIEVEMENTS

Actively participated and led Kalakriti, the official college drama and theatre group, performing at various inter-college festivals
and cultural events.