Integrated Knowledge & AI

Technology In Farming
Capstone Project Report

MID SEMESTER EVALUATION

Submitted by:

101903072 Karan Singh Pathania

101903202 Tarunvir Singh
101903334 Amritpal Singh
101903343 Inderpal Singh

BE Third Year, COE

CPG No: 13

Under the Mentorship of

DR. ANIL SINGH

Assistant Professor

Computer Science and Engineering Department

Thapar Institute of Engineering and Technology, Patiala
July 2022
 ABSTRACT

The use of modern computing technologies like machine learning and artificial intelligence in the
field of agriculture has grown in developed countries but there is a need of low-cost technology
development for farmers in developing countries like India. The major reason behind this growing
popularity and demand is the ability of these machine learning models to accurately predict plant
disease and suggest crop patterns that increase the yield of the land. The question we posed for our
project was whether a holistic solution can be developed which is able to assist the Indian farmers
at a low-cost. We have successfully built a hardware model which is able to collect readings from
the soil used in farms. The analysis from these reading is used to suggest the correct crops and
agricultural practices for the farm which can increase the productivity of the farmland. All this
work is carried out using machine learning models and this can assist the farmers in short as well
as long term. Further we have also successfully trained a machine learning model which can detect
various plant diseases based upon the images uploaded by the farmer in our webapp. Our team’s
IKAT prototype is a very limited version of what could be created in a production facility using
more advanced technology.

While optimizing the existing model we are also working to build a robocar which can successfully
manoeuvre in a farm like terrain and can collect the required readings and click images of the
disease affected crops. Also, our team is currently working implementing the webapp and also the
ripeness detection model. We are trying to build a low-cost system for the small-scale farmers of
developing nation like India.

i
 DECLARATION

We hereby declare that the design principles and working prototype model of the project entitled
Integrated Knowledge & AI Technology in Farming is an authentic record of our own work carried
out in the Computer Science and Engineering Department, TIET, Patiala, under the guidance of
Dr. Anil Singh during 6th semester (2022).

Date:

Roll No. Name Signature

101903072 Karan Singh Pathania
101903202 Tarunvir Singh
101903334 Amritpal Singh
101903343 Inderpal Singh

Counter Signed By:

Faculty Mentor:

Dr. Anil Singh

Assistant Professor
CSED,
TIET, Patiala

ii
 ACKNOWLEDGEMENT

We would like to express our thanks to our mentor Dr. Anil Singh. He has been of great help in
our venture, and an indispensable resource of technical knowledge. He is truly an amazing mentor
to have.

We are also thankful to, Dr Shalini Batara Head, Computer Science and Engineering Department,
entire faculty and staff of Computer Science and Engineering Department, and also our friends
who devoted their valuable time and helped us in all possible ways towards successful completion
of this project. We thank all those who have contributed either directly or indirectly towards this
project.

Lastly, we would also like to thank our families for their unyielding love and encouragement.

They always wanted the best for us and we admire their determination and sacrifice.

Date:

Roll No. Name Signature

101903072 Karan Singh Pathania

101903202 Tarunvir Singh

101903334 Amritpal Singh

101903343 Inderpal Singh

iii
 TABLE OF CONTENTS

ABSTRACT………………………………………………………………………………….i
DECLARATION……………………………………………………………………………ii
ACKNOWLEDGEMENT………………………………………………………………....iii
TABLE OF CONTENTS…………………………………………………………………..iv
LIST OF TABLES………………………………………………………………………....vi
LIST OF FIGURES………………………………………………………………………..vii
LIST OF ABBREVIATIONS……………………………………………………………..viii

CHAPTER………………………………………………………………………………Page No.
1. Introduction 1 1
1.1 Project Overview 1
1.2 Need Analysis 3
1.3 Research Gaps 4
1.4 Problem Definition and Scope 5
1.5 Assumptions and Constraints 5
1.6 Standards 6
1.7 Approved Objectives 6
1.8 Methodology 7
1.9 Project Outcomes and Deliverables 7
1.10 Novelty of Work 8
2. Requirement Analysis 10
2.1 Literature Survey 10
2.1.1 Theory Associated With Problem Area 10
2.1.2 Existing Systems and Solutions 10
2.1.3 Research Findings for Existing Literature 11
2.1.4 Problem Identified 17
2.1.5 Survey of Tools and Technologies Used 18
2.2 Software Requirement Specification 18
2.2.1 Introduction 19
2.2.1.1 Purpose 19
2.2.1.2 Intended Audience and Reading Suggestions 19
2.2.1.3 Project Scope 19
2.2.2 Overall Description 20
2.2.2.1 Product Perspective 20
2.2.2.2 Product Features 20
2.2.3 External Interface Requirements 21
2.2.3.1 User Interfaces 21

iv
 2.2.3.2 Hardware Interfaces 21
2.2.3.3 Software Interfaces 21
2.2.4 Other Non-functional Requirements 22
2.2.4.1 Performance Requirements 22
2.2.4.2 Safety Requirements 22
2.2.4.3 Security Requirements 22
2.3 Cost Analysis 23
2.4 Risk Analysis 24
3. Methodology Adopted 25
3.1 Investigative Techniques 25
3.2 Proposed Solution 25
3.3 Work Breakdown Structure 27
3.4 Tools and Technology 28
4. Design Specifications 29
4.1 System Architecture 29
4.2 Design Level Diagrams 31
4.3 User Interface Diagrams 34
4.4 Snapshots of Working Prototype 36
5. Conclusions and Future Scope 38
5.1 Work Accomplished 38
5.2 Conclusions 38
5.3 Environmental Benefits 39
5.4 Future Work Plan 39
APPENDIX A: References 41
APPENDIX B: Plagiarism Report

v
 LIST OF TABLES

Table No. Caption Page No.

Table 1 Research Finding for Existing Solution 13

Table 2 Cost Analysis 23
Table 3 Investigative Techniques 25

vi
 LIST OF FIGURES

Figure No. Caption Page No.

Figure 3.3.1 Work Breakdown Structure 26
Figure 4.1.1 System Level Architecture 38
Figure 4.1.2 Software Level Architecture 37
Figure 4.2.1 Dataflow diagram 38
Figure 4.2.2 ER diagram 39
Figure 4.3.1 Use case diagram 40
Webpage describing the problem statement and solution
Figure 4.3.1 41
of the project
Figure 4.3.2 Webpage describing our project 41
Figure 4.3.3 Webpage to enter the sensor data 42
Figure 4.3.4 Webpage to upload the image 42
Hardware Circuit connection using
Figure 4.4.1 43
breadboard
Figure 4.4.2 Microprocessor connected to the sensors 44
Figure 4.4.3 Activity showing crop requirement 44
Figure 4.4.4 Activity showing sensor reading 44

vii
 LIST OF ABBREVEATIONS

GPU Graphical Processing Unit

IoT Internet of Things

ML Machine Learning

AI Artificial Intelligence

viii
 INTRODUCTION

1.1 Project Overview

The following section will provide some brief insight on the project.

1.1.1 Technical Terminology

The Integrated Knowledge & AI Technology (IKAT) in Farming is the prominent

technology through which data from several agricultural fields can be collected using
different sensors. The collected data are analyzed by various machine learning models to
draw conclusions based on different climatic patterns, soil fertility, the nature of crops,
and the amount of water that needs to be supplied to the field using IoT devices. This
project mainly focuses on predicting the appropriate crop based on the climatic situations,
predicting the disease of the crop, identifying the weeds in the crop, ripeness of the fruits,
and the yield of the crop based on the historic data by using supervised machine learning
algorithms.

1.1.2 Problem Statement

In Farming, monitoring the huge fields has always been difficult and challenging for the
farmers. As the monitoring the crops include various things such as need to water,
identifying infected crops, checking of ripeness etc. Specifically the difficulty of not
knowing when to water the crops and in what quantity can affect the crops and also lead
to wastage of water. Plants and crops getting affected by diseases and pest has been a very
common problem in farming. Moreover, there is no way to give an exact reference to the
requirements. In the absence of enforceability of contractual provisions, either of the
parties can breach their ends of the contract. Moreover, there is no means to monitor the
crop and guide the farmer during the growing period.

1
1.1.3 Need

Agronomy has a significant contribution in shaping and boosting the economy of

developing nations, especially India. The issues faced by the agricultural zones are
hampering the growth of these economically developing nations. The only remedy to the
crisis is to do all that is possible to make agriculture a profitable enterprise and attract the
farmers to continue the crop production activities. As an effort towards this direction, this
capstone project i.e., IKAT in farming would help the farmers in making appropriate
decisions regarding the cultivations with the help of machine learning.

1.1.4 Solution

In the past farmers used to predict their yield from the previous year's yield experiences.
Thus, for this kind of data analytics in crop prediction, there are different techniques or
algorithms, and with the help of those algorithms, we can predict crop yield. By analyzing
all the issues and problems like weather, temperature, and several factors, there is no
proper solution and technologies to overcome the situation faced. The major benefit of this
proposed work is the implementation of precision agriculture with cloud computing i.e.,
Integrating the Knowledge Management with Artificial Intelligence which will control the
usage of water and monitor other plant parameters (soil moisture, temperature, humidity,
and air quality index) and early prediction of crop diseases/weeds, thus enhancing the yield
of the crops.

1.1.5 Scope of Improvement

I. Crop Monitoring: The various features of soil such as moisture and nutrients
would be continuously monitored and that information would be provided to the
farmer, instead of directly judging the quality of the final product, it will be better
for both the firm and the farmer.

2
 II. Stating the requirements: Stating the growing conditions (like maintaining soil
humidity, pH, fertilizer concentration in a range) instead of verbally stating the
quality requirements of final products will make the contracts more deterministic.

III. Recording violations in an immutable database: Machine learning can be used

in these situations to get correct results, it is very essential to have the knowledge
of when to harvest the crops as it can highly increase the amount yield that is
produced. Thus the farmers will be saved from undue quality cuts and exploitation.

1.2 Need Analysis

Farming is the backbone of the Indian economy and it highly contributes to the GDP of
our country. In India half of the population is involved in agriculture and contributes 7.6%
of total agriculture yield, which is very limited and requires special attention. According
to studies pests, weeds, and diseases cause India to lose 15-25 percent of its potential crop
yield. If the pest and weed are controlled in their initial stages the loss can be reduced.
The need of this project focuses on countering these challenges via continuous crop
monitoring and the deployment of technology such as artificial intelligence. This
technology would deliver vital information to our farmers, allowing them to take essential
safeguards early on to save their crops from damage.
We always have limited amount of water accessible to us for cultivation. Its commonly
noted that the water is not utilized in an efficient way in farming, which leads to its a lot
of wastage. So it becomes a crucial factor to use existing water resources sustainably. For
sustainable use of water in farming there is an essential need to automate the irrigation
which we will do in our project with IOT and AI technologies.
Yields of crops can be increased and a lot of workload can be reduced from the head of
our farmers. This purpose can be fulfilled like:

● In this project, we will assist farmers in determining when the crop is ready to be
harvested by analyzing its ripeness.

3
 ● The crop's productivity would be boosted by detecting critical characteristics such
as soil pH, humidity in the environment, nitrogen/phosphorus content in the soil,
rainfall in the region etc.
● We would also assist farmers in determining which crop would be best to cultivate
in a given region based on the greatest needs anticipated by machine learning.
● A WebApp would provide farmers with easy access to crop monitoring
information.

1.3 Research Gaps

The smart irrigation system that is being used in [2] for purpose of providing water to the
plants through automated techniques essentially uses a setup which is based upon the
concept of centralized technology. In this technology the dysfunction of the controller
would result in whole network failure, to avoid such cases the whole system can be
decentralized. The SVM algorithm used in [3] for the classification of the images that are
being captured by rover would not give higher accuracy, and it also depends which types
of diseases are considered for the process of classification. So better option would be to
use concepts Convolution Neural Network, OpenCV or Xgboost as studies have shown
they give better results in case of images classification. In this paper [4], a system has been
proposed that locates, tracks, and estimates the ripeness and quantity of fruit in the field.
The proposed approach consists of two sub-systems: a detection and ripeness estimation
sub-system and a tracking subsystem, the results given by these respective systems could
be increased using other different combinations of deep learning techniques or combining
the results of different models to push the results in the range of 90 to 100%. In [6] the
analyzing of the characteristics and quality of the soil is done with the hardware system
consisting of various sensors like DHT11 sensor and Soil moisture sensor. Use of DHT11
could affect the results as the sensor is made of plastic, extreme heat can cause it to melt.
Also, due to prolonged exposure to a high-humidity environment, the sensor may
accumulate water vapors due to heat and humidity, which will eventually condense into
liquid inside the covering. To avoid such scenarios the sensors such as DTH22 can be used
which would also further provide better accuracy and measuring range. A Literature Study
on Agriculture Production System Using IOT as Inclusive Technology is done in [8] which

4
basically describes its role in solving problems in domain of agriculture. Although it brings
various essential results, but cost of deploying such technology is higher. Also it has the
problem of integration as there is currently no consensus regarding IOT protocols and
standards. These can be solved by understanding the network and giving any required
customizations like planning for extra time with device deployments to handle any
troubleshooting.

1.4 Project Definition and Scope

Productivity of a farm can be enhanced by determining which crop variety has produced
the greatest yield under similar soil, climate, fertilization, and irrigation conditions. The
same data driven approach to crop selection can also address climate change, resource
constraints and societal concerns around issues such as animal welfare, fertilizers, and
environment that often impact agricultural production. Increased crop productivity is
urgently needed, and it is the cornerstone of any solution for meeting food shortage and
farm profitability problem. Smart farming involves the use of technology and in particular,
the internet of things and related big data analytics to address these challenges via the
electronic monitoring of crops, as well as related environmental, soil, fertilization, and
irrigation conditions. Such monitoring data can be then be analyzed to identify which crops
and specific crop varieties can best meet the productivity targets of any particular farm
around the world. Crop variety identification involves the use of plant phonemics and
related data analysis results with specific crop varieties. The association of information will
revolutionize the way food is produced globally.

1.5 Assumptions and Constraints

Although the factors on which the growth of a crop depends on are many, but w a naming
that it only depends on the most significant four humidity of the air, moisture of the soil,
temperature and wind.
The sheets on the outer of the Greenhouse can still be damaged if there is a strong
windstorm or rain. We are assuming that weather conditions do not vary to that extent.

5
The constraints for the project include the following
● Arduino Uno Board has only 13 input/output ports, out of which 2 are used for
Receiver-Transfer logic; so, we are only lett with 11 ports. Therefore, we must
mold the logics such that at maximum, we use 11 ports.

● We cannot put our model in an open field, since our model is not that much
impregnable that it would withhold and withstand rains and strong winds or hails.
In addition, there is always a danger of animals and birds too.

1.6 Standards

Software System
• Tools: Web browser, Arduino IDE.
• Language: Python (version 3.6.5), Javascript (ES6) etc.
• Plugins & libraries: Primarily inbuilt Arduino plugin and libraries, Tensorflow, cupy,
cuml library etc.
• Techonolgy Used: Machine Learning, Deep Learning and Data-preprocessing etc.

Hardware System
• Sensor: soil moisture sensor, temperature/humidity sensor, rain detector sensor and PIR
sensor.
• Data Processing: Processor used is nodemcu (ESP 8266).
• Device: Smart Phone/Tablet.

1.7 Approved Objectives

● The irrigation system which is used to water the plants is planned to be automated, in
which the water would be supplied according to the readings taken from the sensors.
● Disease is automatically detected in the crops through the images of their leaves and it
would be passed to the AI model to classify the type of disease.

6
● All the aspects that would be required to check the conditions of the crops are measured
through sensors. The parameters that would be checked are humidity, temperature etc.
● The crop prediction would be done on the basis of the region where we have to do the
farming. And the best type of crop suitable to grow in that region would be shown on
the website.
● All the systems are connected to the website and the user is provided with the master
control of all the systems.
● All the measurements or the readings related to crops would be provided on the website.
These reading will tell about the soil conditions.

1.8 Methodology

We achieve these tasks for projects by firstly, implementing the hardware design of the
embedded system we are using. This embedded system, to be installed at some elevation
in the fields, would contain all the sensors, including wind current sensor, temperature and
rain detector. These statistics would keep on getting updated on the online database that
could be accessed by the owner of the field on his android application.

In the Robo car, a human controls the car by remote control. This is done by sending and
receiving the signals from the car. The car can thus cover the whole farm and capture the
picture of plants which can further be sorted by machine learning that which crops are
having any disease. The receiver has been configured to take input coordinates from the
web app designed to make it autonomous.
Raspberry Pi will be used to analyze the crop field and provide possible points to check in
the agriculture field. The webapp would be made that would have the master control over
the whole control system. The user can also see the database with this application so one
can evaluate the conditions over his field.

1.9 Project Outcomes and Deliverables

Our project is based on both hardware as well as software implementation.

7
 Physical Model:

Our project is to be built on a physical model. All the conditions as well as the
second outcome, i.e. webapp is also based on this. The model would contain all
sensors that would be deployed to check the weather conditions of the soil. And on
the basis of that there would be certain actuators which would regulate the weather
system so that it is favorable for the plant. At the successful completion of the
project a fully autonomous robot that can move in a farm environment without
damaging existing plants or soil and use object detection to find and mark diseased
crops with an environmentally safe color. And all the data of the conditions would
be uploaded over to an online database through the Arduino and web app. This will
help farmers in producing a good yield and also prevent them from any kind of
loss of seeds etc. which get rotten if they are given an amount of water higher than
their requirement.

Web App:

Our second outcome is the app for the integrated knowledge & AI technology in
farming. This site would be able to access the database that was used to upload all
the weather data by the model. Thus, able to check all the statistics, the user would
be able to analyze the crop growth so as to control the condition ranges that are to
be inputted by the user. This webapp would thus have to be built such that it gives
master control over all the sensors of the physical model, and their ranges.

1.10 Novelty

Our Project is different from other available products in market as follows

● Cost: The total cost of implementing the microcontrollers and sensors very less
as compared to other products available in market. Hence, ever mall scale farmers
can implement that in their small fields

8
● Scalability: The product is highly scalable, as the number of users inrease there
will be very small modification needed to implement the system.

● Open Source Application: The Android application used is open source and the
whole source code is available on Github. Hence, the users can tweak the
application specifically for their needs and requirements.

● Ease of Use: The product is very easy to use. Just install the sensors in the
Gremhouse and connect them to specified ports in Microcontroller. Then we just
need to install Android application to use the system.

9
 REQUIREMENT ANALYSIS

2.1 Literature Survey

The following chapter covers the details about the literature survey done, discussing about
the existing technology, the technology on which the team worked and theory associated
to the problem area.

2.1.1 Theory Associated with Problem Area

Farming is the backbone of the Indian economy and it highly contributes to the GDP of our
country. In India half of the population is involved in agriculture and contributes 7.6% of
total agriculture yield, which is very limited and requires special attention. According to
studies pests, weeds, and diseases cause India to lose 15-25 percent of its potential crop
yield. If the pest and weed are controlled in their initial stages the loss can be reduced.
Agriculture faces a number of unique social, economic and environmental challenges,
including increasing globalization and international competitiveness, climate variability,
shortages in labour, urban pressure on farmland. Increasing population requires food
production to be increased to meet the demand, which requires better cultivation in the
form of proper utilization of seeds and fertilizers with minimum labour work. Use of
advanced techniques like making use of machine learning and artificial intelligence to
predict the nature of crop to be grown, to detect crop diseases and to check the ripeness of
the fruits can reduce the dependence on manual labour and increased the yield obtained at
the same time. Though much advancement in the field of agriculture has reduced the
reliability on manual labour, there is a strong need use advance computer technologies for
increasing the productivity of the farmland.

2.1.2 Existing System and Solutions

Following are some of the major players that are operating in the market.
• Plain Sight is a USA based vision AI platform that enables livestock and crop
producers and processors to increase efficiency and generate more revenue .

10
 ● Gro Intelligence is a USA based analytic platform that provide a unified view
of climate, agriculture and economy insights. They help in managing the food
security and help farmers in making short term and long term decisions,
● Dickey John is a USA based organization that works in a lot of domains and
provide a lot of smart products and sensors that can help in better monitoring of
crops.
● Gamaya is a Swiss based AgTech start up that uses cutting-edge system for
mapping and diagnosing farmland to address the need to boost the productivity
and sustainability of the agricultural industry. Their product mainly focusses on
using drones to diagnose the farmland and optimize the farming area to provide
better re-planting decisions.
● Farmers Edge is a Canadian based group that analyses existing crop data and
helps in better decision making by suggesting seeds , manure and other tools
required for farming.

2.1.3 Research Findings for Existing Solution

S.No Roll No. Name Paper Title Findings Citations

Reliable Data Fusion of
data fusion problem
Hierarchical Wireless
of wireless sensor
Karan Singh Sensor Networks with
1 101903072 networks for the [1]
Pathania Asynchronous
green house
Measurement for
monitoring system
Greenhouse Monitoring
Smart Irrigation System: Water Conservation
IOSR Journal of by automatically
Karan Singh
2 101903072 Electronics and providing water to [2]
Pathania
Communication plants based on
Engineering needs.

11
 An autonomous rover
Smart Irrigation with Field that monitors the
Protection and Crop Health field and, a system
3 101903202 Tarunvir Singh [3]
Monitoring system using for identification and
Autonomous Rover classification of
affected plants
A system that locates,
Fruit Quantity and tracks, and estimates
4 101903202 Tarunvir Singh Ripeness Estimation Using the ripeness and [4]
a Robotic Vision System quantity of fruit in
the field
ML-ASC
performance
framework proposed
in the study can
Smart Farming Using
5 101903334 Amritpal Singh transform the existing [5]
Machine Learning
production systems
into data-driven
smart manufacturing
systems.
a software
Special Issue of Second application that
International Conference provides suggestions
6 101903334 Amritpal Singh [6]
on Advancements in to farmers when the
Research and Development hardware system
analyses the soil
agricultural
Robotic applications in the
automation under
automation of agricultural
7 101903343 Inderpal Singh greenhouse through [7]
production under green
the use of robotics
house: A review
platforms as a source

12
 to solution to
different problems
that arise in the
agricultural sector
data collection,
network service, date
fusion and
computation,
A Literature Study on
analysing the
Agriculture Production
8 101903343 Inderpal Singh technologies, based [8]
System Using IOT as
on which further the
Inclusive Technology
technological system
framework of the
internet of things are
brought forth

Table 1 : Research Findings for Existing Solution

• Reliable Data Fusion of Hierarchical Wireless Sensor Networks with Asynchronous

Measurement for Greenhouse Monitoring [1]
This paper investigates the data fusion problem of wireless sensor networks for the green
house monitoring system. Considering the characteristics of local consistency and slow
change of the greenhouse environmental information, the hierarchical structure of WSNs
is proposed for the greenhouse monitoring system, and the two-stage data fusion scheme
is presented for the hierarchical network. The multi-rate measurement mode is proposed to
reduce the energy consumption of WSNs under the premise of satisfying the information
sensing performance of the system. The data fusion at the sink node is conducted on the
support function with the consistency analysis of data from different clusters.

13
• Smart Irrigation System: IOSR Journal of Electronics and Communication
Engineering (IOSR-JECE) [2]
With the water requirements in irrigation being large, there is a need for a smart irrigation
system that can save about 80% of the water. This prototype aims at saving time and
avoiding problems like constant vigilance. It also helps in water conservation by
automatically providing water to the plants/ gardens depending on their water
requirements. It can also prove to be efficient in Agricultural fields, Lawns & Parks. As
technology is advancing, there is always a chance of reducing risks and making work
simpler. Embedded and microcontroller systems provide solutions for many problems.
This application precisely controls the water system for gardens by using a sensor
microcontroller system. It is achieved by installing sensors in the field to monitor the soil
temperature and soil moisture, which transmits the data to the microcontroller for
estimation of water demands of plants.

• Smart Irrigation with Field Protection and Crop Health Monitoring system using
Autonomous Rover [3]
In this research paper work has been done on an autonomous rover that monitors the field
and, a system for identification and classification of affected plants. For the purpose of
monitored and sustainable irrigation, a DTH11 module and a water level sensor module
has been combined to draw the conclusion whether the crop can be watered or not. When
the water level in the soil is below the required water level essential for the plant, the
temperature and humidity are checked. This is done considering the necessity of water
during inevitable plight. Classification of disease is done using a support vector machine
algorithm, which helps in identifying the type of disease that has affected the plant. Also,
when the farmer wishes to check his land personally, the rover may be manually operated
using IoT and accordingly take suitable actions. A smart autonomous gardening rover,
which can be controlled remotely has been fabricated which automatically identifies,
classifies the plant species and measures the key parameters for farming such as
temperature, humidity, heat level, wind speed, wind direction and soil moisture. Based on
the acquired data and history, actions are taken to maintain the farm more effectively.

14
• Fruit Quantity and Ripeness Estimation Using a Robotic Vision System [4]
In this paper they have proposed a system that locates, tracks, and estimates the ripeness
and quantity of fruit in the field. The proposed approach consists of two sub-systems: a
detection and ripeness estimation sub-system and a tracking subsystem. The detection and
ripeness estimation sub-system uses the Deep Fruits technique and proposes a new network
structure that jointly learns detection and ripeness estimation. With Deep Fruits, they have
fine-tuned a pre-trained FRCNN network. Tracking via detection provides a simple
framework to count the number of fruits that can be deployed using off-the-shelf cameras
without the need for three-dimensional, inertia, or odometry information. Overall, in their
first approach, they pose a multi-class task where each ripeness is its own class to be
detected, referred to as Multi Class-FRCNN. The second approach poses the task as two
parallel layers (Parallel-FRCNN): one layer to perform detection; and a second to estimate
ripeness. Evaluation of these two approaches outlines the improved performance of the
parallel layer, where they achieve an F1 score of 77.3 for the parallel technique yet only
72.5 for the best scoring multi class implementation.

• Smart Farming Using Machine Learning [5]

The study focuses on ML applications in analysing the data, and less attention is paid on
how the data from different innovative technologies is captured, stored, analysed, and
shared across the different phases in ASC. More research is required in this direction. The
findings indicate that ML helps to enhance ASC visibility. However, it would be interesting
to know the impact of such influence. The future studies should aim at measuring the
impact of ML on ASC and provide specific guidelines on how the ASCs should deploy
ML for enhanced supply chain visibility. The ML-ASC performance framework proposed
in the study may be used as a guiding framework for such studies. The future studies should
focus on how ML can transform the existing production systems into data-driven smart
manufacturing systems. The future studies should aim at developing focused customer
frameworks, which should capture insights on customer buying behaviour.

15
• Special Issue of Second International Conference on Advancements in Research and
Development (ICARD 2021) [6]
The smart Farming using IOT system mentioned in this research paper is a hardware
integrated with a software application that provides suggestions to farmers when the
hardware system analyses the soil. After analysing the characteristics and quality of the
soil with the hardware system consisting of various sensors like DHT11 sensor and Soil
moisture sensor, it provides the analysed data to the application via the Internet. Then this
application compares the data in the database and provides user suggestions and also
remotely monitors and controls equipment like drip irrigation systems and electric fencing,
etc. The databases system analyses the data provided by the hardware as input and gives
the user suggestions like, which crop is best suited for this soil, its organic farming methods
and irrigation methods, etc. After that, it can also predict any animal intrusion by using a
PIR sensor. This application mainly uses data analytics and database management
techniques to derive suitable crops and their cultivation methods from the data sets that
were collected from the research centres and organic farmers. And also helps to monitor
and control farmland using IoT.

• Robotic applications in the automation of agricultural production under green house:

A review [7]
This article shows a review of the state of the art in the agricultural automation under
greenhouse through the use of robotics platforms as a source to solution to different
problems that arise in the agricultural sector specially under controlled conditions; the
writing is part of a wide review of the agricultural automation under greenhouse and a set
of proposals that consider applications in various fields of research, autonomous localized
irrigation, neural networks, automatic learning, monitoring and control systems among
others. It is intended to determine the most recent advances in techniques for sowing,
applications of chemicals for sanitary control, fertilization, harvest, where robotics is used
as a source to solve agricultural problems. The review includes consultation in well-known
national and international journals, research carried out in centres and research groups or
companies dedicated to implementation addressed to the agricultural sector, finally it is

16
 determined the solutions that can be implemented to propose new topics, research lines,
development and technology transfer to the agricultural sector.

• A Literature Study on Agriculture Production System Using IOT as Inclusive

Technology [8]
Directing at the current development condition of the internet of things and based on the
available technology analysis of the internet of things, analysis and research on the internet
of things in terms of technological levels and systems are made. Started from three aspects,
respectively, data collection, network service, date fusion and computation, analysing the
technologies, based on which further the technological system framework of the internet
of things are brought forth. Moreover, analysis and research work on the sensor nodes of
the system, analysis and discussion on the various technologies involved are carried on.
Based on the current IOT technology analysis, by analysing and discussing the
technological levels and systems of IOT, it is going to start the research on the architecture
and the frame work of IOT. Started from intelligent transportation, logistic scheduling and
tracing and base station monitoring, IOT extends its application domain to public oriented
personal medical treatment, intelligent home furnishing and so on, and its applications can
be found in all walks of life.

2.1.4 Problems Identified

We analysed all the solutions available in the market and concluded that there is very less
Indian based solution available. Since India has a unique climate and soil features which
are different from the rest of the countries there is a dire need of AgTech based start-ups in
India which focus primarily on the Indian Market and help in increasing the productivity
of the farmers. Further there are very less systems available yet which automate the task of
pest detection, ripeness detections, moisture control, humidity control and analyse the data
to provide short term and long-term solutions to the farmers that increase the productivity
of the farmland. The available systems are either for a specific purpose or are expensive
hence a small-scale Indian farmer is not able to afford them. We are aiming to build a
holistic solution which primarily targets the Indian Market.

17
2.1.5 Survey of Tools and Technologies Used
Following are the components that will be used in Integrated Knowledge & AI
Technology in Farming in no particular sequence
Hardware Parts Used: -
● Nodemcu Esp8266
● Soil Moisture Sensor
● Temperature Sensor
● Humidity Sensor
● Mini Water Pump
● Raindrop Sensor
● PIR Sensor
● Arduino UNO
● Breadboard
Software Used: -
● Tensor Flow
● Anaconda
● Visual Studio Code
● Mongo DB Atlas
● Postman
● MS Word
● MS Power Point

2.2 Software Requirement Specification

The following sections describe the Software requirement specification for this project.

18
2.2.1 Introduction

2.2.1.1 Purpose
The website is going to be developed and all the sensors and actuators will be controlled
by Nodemcu whose programming will be done in Arduino IDE. ML model will be
developed for detecting the disease of the crops.

2.2.1.2 Intending Audience and Reading Suggestions

The intended use of this project revolves around the farmers for helping them to connect
with a new technology which can be useful in a way that can produce a good yield of the
crop and also reduce their time wastage and loss.

2.2.1.3 Project Scope

This project aims towards developing a website and blynk app-controlled Intelligent
System, which will automate the water pump and will help provide a better user experience
by providing users with all the data in real-time. On a brief account, the website will be
used on the phone of the farmer, so that they would be able to irrigate their farms over to
the internet and the website will also be used to monitor the data.

It would be easy to use and easy to carry. The disease can be detected using a pre-trained
machine learning model. The project focuses on providing the farmers with something
completely different that will save them the hard job of reaching every corner of the field
to check for some abnormality.

19
2.2.2 Overall Description

2.2.2.1 Product Perspective

Agriculture employs more than 50% of the Indian workforce. In addition, it is the backbone
of our GDP. The Intelligent Farming system aims to automate the process of irrigation and
disease detection. With our project, we will bring the power of computing to the Indian
Farmer. We would reduce the need of being physically present on the farm. We are making
a smart irrigation system, which automatically irrigates the fields of farms by checking the
moisturizer in the soil also our project will irrigate, the fields by checking the weather
conditions. With this greenhouse environment shielding and temperature control system,
etc. will all be automated. The farmer would be able to control these tasks through his
smartphone from anywhere around the world. He would be able to see the statistics and
warnings related to the farm on his smartphone. By this, our system can boost the
productivity of the same piece of land, as all the resources will be micromanaged by the
software system.

2.2.2.2 Product Features

The software will be able to perform the following operations:

1. Live Notification: Live feedback to the farmer to check the abnormality in the field (if
present).

2. Manual Mode Operations: When logged in to the manual mode, the farmer will be
allowed to change the state of any actuator without any consideration for the sensor
data. For example, the farmer can turn on/off any of the motor, sprinkler, etc. manually
without sensor data interruption.

20
3. Show Sensor Data: The data from the sensors will be displayed in this mode in real-
time. Data will be collected from sensors including temp/humidity sensor, soil
moisture sensor, rain sensor, etc. Initially, we plan to implement these functionalities
for one crop under that as part of the Pilot Phase. Once the Pilot Phase is successful,
we plan to spread this application across multiple crops.

2.2.3 External Interface Requirements

The External Interface Requirements are defined in the following sub-sections.

2.2.3.1 User Interfaces

An android device/ Pc would be used for this purpose to fetch the current values of sensors
and actuators and it would display all of them and the statistics of all the crop data. A
website solely made for this purpose is made.

2.2.3.2 Hardware Interfaces

Devices: Smart Phones or Tablets, Nodemcu
Sensors or Modules: Soil Moisture Sensor, Temp/Humidity Sensor, Rain Sensor, PIR
sensor, Water Pump

2.2.3.3 Software Interfaces

Operating System: Android (Latest), Windows
Tools: Android studio, Arduino IDE
Language: Python
Plugins & libraries: Primarily inbuilt android and Arduino plugins and Libraries

21
2.2.4 Other Non- functional Requirements

Following are some of the non-functional requirements of the project.

2.2.4.1 Performance Requirements

● The system is designed for smartphones/Pc and hence requires a smartphone running
android or windows pc to access the system.
● Only textual information will be handled by the website. The amount of information
to be handled can vary from user to user.
● For normal conditions, automatic mode will work fine. If the user wants to access
manual mode, he can do that too.

2.2.4.2 Safety and Security Requirements

● The source code developed for this system shall be maintained in the configuration
management tool.
● The whole system is secured. Only the registered farmers can access the data.
● Moisture sensor pens should be taken care of because they are very brittle.
● There will be no issue of loss or theft of user personal data as our device requires no
such information.

22
2.3 Cost Analysis

List of all the components along with their cost that will be used in the project:

S. No. Component Price

1. Nodemcu Esp8266 350/-

2. Soil Moisture Sensor 180/-

3. Temp-Humidity Sensor 260/-

4. 5V Relay One Channel 150/-

5. Jumper Wires 240/-

6. Mini Water Pump 250

7. Raindrop Sensor 200/-

8. PIR Sensor 100/-

9. Breadboard 160/-

10. Battery 60/-

TOTAL 1950/-

Table 2 : Cost Analysis

Total cost for one unit is Rs 1950/-

23
2.4 Risk Analysis

Four Things to Consider Before Developing Your Smart Farming Solution:

As we can see, the use cases for IoT in agriculture are endless. There are many ways smart
devices can help you increase your farm's performance and revenue. However, agriculture
IoT web development is no easy task. There are certain challenges you need to be aware
of if you are considering investing in smart farming.

1. Hardware
To build an IoT solution for agriculture, you need to choose the sensors for your device
(or create a custom one). Your choice will depend on the types of information you want
to collect and the purpose of your solution. In any case, the quality of your sensors is
crucial to the success of your product- it will depend on the accuracy of the collected
data and its reliability.

2. The Brain
Data analytics should be at the core of every smart agriculture solution. The collected
data itself will be of little help if you cannot make sense of it. Thus, you need to have
powerful data analytics capabilities and apply predictive algorithms and machine
learning to obtain actionable insights based on the collected data.

3. Maintenance
Maintenance of your hardware is a challenge that is of primary importance for IT
products in agriculture, as the sensors are typically used in the field and can be easily
damaged. Thus, you need to make sure your hardware is durable and easy to maintain.
Otherwise, you will need to replace your sensors more often than
you would like.

24
 METHODOLOGIES ADOPTED

3.1 Investigative Techniques

Technology used in Investigative

S.No Investigative Description
project Techniques

Sensor based Soil moisture, water & light management,

1 Descriptive
technology humidity & temperature control

Software based Machine Learning Methods for appropriate

2 Comparative
technology medicine recomendor for plant disease
Deep Learning- Plant Disease Recognition, Artificial
3 Comparative
based technology Intelligence in Crop Yield Prediction
Decision making and prediction processes
Data analysis related
4 Experimental and managing all crop and weather
technology
information
Hardware based Capturing images of plants using camera with
5 Experimental
technology all sensors attached to carbot

Table 3: Investigative Techniques

3.2 Proposed Solution

We achieve these tasks for the project by implementing the hardware design of the
embedded system we are using. This embedded system, to be installed at some elevation
in the fields, would contain all the sensors including a soil moisture sensor, temperature
and humidity sensor, PIR sensor and rain detector. These statistics would keep on getting
updated on the online Blynk platform that could be accessed by the owner of the field on
his Blynk application.

Also, this embedded system would be connected to the water pump which can be switched
on from the blynk platform itself whenever the soil moisture is low. Mobile notifications

25
will come on the user's application which would be indicating low moisture levels, or
whenever rain is there in the field, about watering your field indication or whenever some
motion is detected in the field. We will use Nodemcu to make this whole control system in
the field.

One can use the pictures of diseased crops/plants and use these pictures to check if there is
an outbreak of the disease in any part of the field using a Machine Learning Model which
is trained well using various algorithms for the best accuracy.

The website would be made that would have the master control over the whole control
system. The user can also see the current readings with the blynk application so he can
currently evaluate the conditions over his field and can alter anything he wants to on the
field.

26
3.3 Work Breakdown Structure

Figure 3.3.1 Work breakdown structure

27
3.4 Tools And Technology
● Hardware used- Nodemcu (ESP 8266), Soil moisture sensor, Temperature and
humidity sensor, Rain Detector sensor, PIR sensor, Water pump
● Software used- Arduino IDE, Python
● Libraries used- Tensorflow
● Technologies used- Machine Learning and data pre-processing.

28
 DESIGN SPECIFICATIONS

4.1 System Architecture

Figure 4.1.1 System Level Architecture

A system architecture, shown in Figure 4.1.1, is a conceptual model that describes the
composition, operation, and other aspects of the system. This architecture is a formal
description and representation of a system that is set up to facilitate analysis of its
components and actions. It consists of developed subsystems and system components that
will interact to construct the overall system.

29
 Figure 4.1.2 Software Level Architecture

A software architecture, as illustrated in Figure 4.1.2, outlines the components that make
up a software system as well as the methodology used to develop them. Each structure is
made up of software elements, their connections, and the attributes of both the elements
and the connections.

30
4.2 Design Level Diagram

Figure 4.2.1 Dataflow diagram

Figure 4.2.1 shows Data Flow Diagram. It is a type of flowchart that illustrates how to
design or debug relational databases in the fields of software engineering, business
information systems, education and research.

31
 Figure 4.2.2 ER diagram

Figure 4.2.2 shows Entity Relationship (ER) Diagram. It is a type of flowchart that
illustrates how "entities" such as people, objects or concepts relate to each other within a
system. ER Diagrams are most often used to design or debug relational databases in the
fields of software engineering, business information systems, education and research.

32
 Figure 4.2.3 Use case diagram

Use-case diagram help to capture the requirements of the system by modelling the
behaviour of the system. It gives the system's high-level operations and domain. The
diagram show’s how the user interact with the system. Use-case diagrams' use cases and
user describe what the system does and how the user use it, but they do not explain how
the system functions internally. The context and requirements of either the entire system
and the key components of the system are illustrated and defined by use-case diagram.

33
4.3 User Interface Diagram

Figure 4.3.1 Webpage describing the problem statement and solution

of the project

Figure 4.3.2 Webpage describing our project

34
Figure 4.3.3 Webpage to enter the sensor data

Figure 4.3.4 Webpage to upload the image

35
4.4 Snapshots of Working Prototype

Figure 4.4.1 Hardware Circuit connection using breadboard

Figure 4.4.2 Microprocessor connected to the sensors

36
Figure 4.4.3 Activity showing crop Figure 4.4.4 Activity showing sensor
requirement reading

37
 CONCLUSION AND FUTURE SCOPE

5.1 Work Accomplished

The question we posed for our project was whether a holistic solution can be developed
which is able to assist the Indian farmers at a low-cost. We have successfully built a
hardware model which is able to collect readings from the soil used in farms. The analysis
from these reading is used to suggest the correct crops and agricultural practices for the
farm which can increase the productivity of the farmland. All this work is carried out using
machine learning models and this can assist the farmers in short as well as long term.
Further we have also successfully trained a machine learning model which can detect
various plant diseases based upon the images uploaded by the farmer in our webapp. The
User Interface of our web app has also been successfully built.

5.2 Conclusion
Integrated Knowledge & AI Technology in farming is a holistic solution for the problems
faced by the farmers of developing nations like India. It is a cost-effective way of smart
farming which can assist the farmer and can increase the productivity of the farmland. We
as a team our trying to build a hardware and a software service to assist the farmers.
Farmers can use our hardware model to analyse the various parameters regarding their
crops like temperature, moisture, humidity, Ph values etc. Now this information once
passed from our machine learning model can suggest the right crop and farming techniques
for the farmers to increase the productivity of their land. Further we are trying to implement
a system where farmer can upload the images of their crop and unwanted crops or crop
diseases can be detected. Also, it is often observed that most of the fruits or vegetables that
come in the market are unripe. Due to this farmer get paid lees for the hard work that they
have put during the whole season. We are also trying to develop a system where we can
detect whether the crop has ripped or not. This project can undergo further research to
improve the functionality of device and its applicable areas.

38
5.3 Environmental / Social / Economic Benefits
Environmental: The second, larger-scale benefit of targeting inputs-in spatial, temporal
and quantitative terms-concerns environmental impacts. The most important
environmental factors on a particular location are sun, air, soil and water. Of the four, water
and soil quality and quantity are the most controllable by humans through time and labour.
Though sunlight & air are available in abundance, crops and plants also depend on soil
nutrients and the availability of freshwater. When farmers are involved in various stages of
farming, sometimes, unknowingly they remove some of these nutrients from the soil. In
future, Internet based farming can significantly reduce the amount of nutrient and other
crop inputs used for boosting the yields, as natural inputs would be sufficient. Smart
farming focuses on replenishing the soil quality while minimizing the use of unnecessary
synthetic resources. Our system will help in saving water and not degrading the quality of
soil due to over irrigation.

Economic: The adoption of our technology will result in several benefits for the farmer
with regard to the economic outcome of his enterprise. A more efficient use of our product
will lead to lower costs and a better quality and production rate. Farmers would be able
to get insights about the crop and their land and would be able to increase the productivity.
Better planning tools might reduce machine operations. Hence, costs for energy and water
can be saved by smart metering.

Social: With technology being so easily available, more and more people could benefit
from it as a whole to improve their day to day lives. This would lead to betterment of their
life comfort, personal space etc leading to overall better life style. A solution provided like
this helps farmers gain their confidence back. Thus, it would have positive social effects in
the forthcoming years.

5.4 Future Work Plan

Our primary focus is to train our and implement our machine learning models and to further
increase the accuracy of the models which we have implemented.

39
Our next task is to successfully integrate all these models on our application so that it can
be made available for the farmers.
Currently the farmer has to manually upload the images and sensory values on our web app
to get the result, but we would like to build a robot – car which could collect all these
information’s and click images of the farmland and provide the insights about the farm
using Internet of Things.

40
 REFRECENCES

[1] Bai X, Wang Z, Sheng L, Wang Z. Sensor Networks With Asynchronous

Measurement for Greenhouse Monitoring. IEEE Trans Control Syst Technol
2019;27:1036–46.
[2] Parthasarathy S, Arun T, Hariharan S, Lakshmanan D. Smart irrigation system. Int
J Innov Technol Explor Eng 2019;8:580–5. https://doi.org/10.48175/ijarsct-2279.
[3] Gobhinath S, Devi Darshini M, Durga K, Hari Priyanga R. Smart Irrigation with
Field Protection and Crop Health Monitoring system using Autonomous Rover.
2019 5th Int Conf Adv Comput Commun Syst ICACCS 2019 2019:198–203.
https://doi.org/10.1109/ICACCS.2019.8728468.
[4] Halstead M, McCool C, Denman S, Perez T, Fookes C. Fruit Quantity and Ripeness
Estimation Using a Robotic Vision System. IEEE Robot Autom Lett 2018;3:2995–
3002. https://doi.org/10.1109/LRA.2018.2849514.
[5] Prof. Swati Dhabarde, Swapnil Bisane, Arti Yadav, Devyani Pote, Akshay Gupta.
Smart Farming Using Machine Learning. Int J Adv Res Sci Commun Technol
2022:140–7. https://doi.org/10.48175/ijarsct-2615.
[6] Shankar V, Kandagatla S, Dorthi K, Rao VCS, Ankam P, Swathi N. Smart farming
using IOT. AIP Conf Proc 2022;2418:63–7. https://doi.org/10.1063/5.0082563.
[7] Tangarife HI, Diaz AE. Robotic applications in the automation of agricultural
production under greenhouse: A review. 2017 IEEE 3rd Colomb Conf Autom
Control CCAC 2017 - Conf Proc 2018;2018-January:1–6.
https://doi.org/10.1109/CCAC.2017.8276478.
[8] K C. A Literature Study on Agricultural Production System Using IoT as Inclusive
Technology. Int J Innov Technol Res 2016;4:2727–31.

41