VISVESVARAYA TECHNOLOGICAL UNIVERSITY

“Jnana Sangama”, Belagavi – 590018, Karnataka.

A
Mini-Project Report
On
“IOT based crop monitoring and ML based crop
recommendation”
Submitted in partial fulfillment for the award of degree of
BACHELOR OF ENGINEERING
In
ELECTRONICS & COMMUNICATION ENGINEERING

Project Associates

NAME USN
1. ABHISHEK H S 4GM20EC002
2. KALLESHA G A 4GM20EC043
3. AISHWARYA M C 4GM21EC402
4. DARSHAN P S 4GM21EC406

For the Academic Year 2023-24

Under the Guidance of
Mrs. Manjula B K
Assistant Professor

Dr. Rajashekhar Somasagar Dr. Praveen J

Project Coordinator Head of the Department

DEPARTMENT OF ELECTRONICS AND COMMUNICATION

ENGINEERING
 GM INSTITUTE OF TECHNOLOGY
Department of Electronics & Communication Engineering
Post Box No. 4, P.B. Road, Davangere – 577 006, Karnataka State.
Email: [email protected] Website: www.gmit.ac.in

CERTIFICATES
This is to certified that, the mini-project entitled “IOT based crop monitoring and ML based
crop recommendation” carried out by Abhishek H S (4GM20EC002), Kallesha G A
(4GM20EC043), Aishwarya M C (4GM21EC402), Darshan P S (4GM21EC406) bonafide
students of VI semester in partial fulfillment for the award of Bachelor of Engineering in
ELECTRONICS & COMMUNICATION ENGINEERING of the Visvesvaraya
University, Belagavi for the academic year2022-23.

Mrs. Manjula B K Dr. Rajashekhar Somasagar

Project Guide Project Coordinator

Dr. Praveen J
Head of the Department

Name of Examiners Signature with date

1. 1.

2. 2.
 ACKNOWLEDGEMENT

The joy and satisfaction that accompany the successful completion of any task would be
incomplete without the mention of the people who made it possible.
We would like to express our gratitude to the Principal, Dr. Sanjay Pande M B, for
providing us a congenial environment for engineering studies and also for having showed us the
way to carry out the project.
We consider it a privilege and honor to express our sincere thanks to, Dr. Praveen J,
IQAC-Director, Professor and Head, Department of Electronics & Communication Engineering
for his support and invaluable guidance throughout the tenure of this project.
We would like to thank our Guide Mrs. Manjula B K Assistant Professor Department
of Electronics & Communication Engineering for support, guidance, motivation, encouragement
forthe successful completion of this project.
We would also like to thank Mini-Project Coordinators Dr. Rajashekhar somasagar,
Associate Professor and Dr. Kavitha K J, Associate Professor, Department of Electronics &
Communication Engineering for their continual support and guidance from the scratch and
throughout the process of project.
We intend to thank all the teaching and non-teaching staff of Department of
Electronics & Communication Engineering for their immense help and co-operation.

Finally, we would like to express our gratitude to parents and friends who always stood
by us.

Project Associates
1. ABHISHEK H S (4GM20EC002)
2. KALLESH G A (4GM20EC043)
3. AISHWARYA M C (4GM21EC402)
4. DARSHAN P S (4GM21EC406)
 ABSTRACT

The process of analysing and extracting valuable information from vast volumes of data is known
as data mining. Data mining is used in many different industries, including finance, retail,
medicine, agriculture, and more. In order to analyse the various biotic and abiotic factors that
influence the crops, data mining is employed in agriculture. The Indian economy's most significant
industry is agriculture, which also offers the bulk of work possibilities. The failure of Indian
farmers to select the right crop for their soil conditions is one of their most common issues. They
are thus observing a sharp decline in productivity. Farmers now have a solution to this problem
thanks to precision agriculture. Precision farming Using research on soil characteristics, soil types,
and crop yield statistics, precision agriculture is a modern agricultural technique that advises
farmers on the best crop to cultivate based on site-specific criteria. More people are interested in
precision agriculture.
It is often proved farmers are the back bone of the country, the agricultural produce not only
helps save life of people it has great role in economic growth of the country. We have been
witnessing role of technology in our day-to-day life, it has completely changed the way we live,
think and do our routines. The same technology could be used to better the life of framer in
managing their task. The proposed project is one such attempt to help farmers manage the
irrigation of their field.
In conventional system they have to visit the field based on the various parameters and
crops irrigation must be done, if field is far away from house farmers find it difficult to manage
these chores, scarcity of electricity is one more menace. To deal with these problems proposed
project provides IOT based solution, here sensors are deployed at fields, these sensors sends
information to base stations powered by ESP32 module, this module uploads all the info to cloud
server which is visible to farmers by means of app, using app they could monitor the status of
irrigation and control the water pump from remote place.
 CONTENTS

Ch. Page
No. Description No.

Acknowledgement
Abstract i
List of Figures Ii
1 Introduction 1-9
1.1 Overview 7
1.2 Problem statement 7
1.3 Objectives 8
1.4 Motivation of the Project 8
2 Literature Review 10-12
3 Methodology 13-34
4 Hardware and Software Requirements 35-40
4.1 Hardware Requirements 35-38
4.2 Software Requirements 38-40
5 Results and Discussion 41-42
6 Advantages, Disadvantages and Applications 43-44
7.1 Advantages 43
7.2 Disadvantages 43
7.3 Applications 43

8 Conclusion and Future Work 44

References
 LIST OF
FIGURES

Figure. Page
No. Description No

3.2.1 System architecture of the purposed system 14

3.3.1 The systems used case diagram is shown above 17
3.4.1 use case diagram of ensemble using KNN module 18
3.4.2 use case diagram of ensemble using naïve bayes module 19
3.5.1 Data flow diagram of system 20
3.5.1.1 Data flow diagram of KNN 21
3.5.2.1 Data flow diagram of Naïve bayes module 22
3.6.1 State chat diagram of system 23
3.6.1.1 State chart diagram of KNN algorithm used for crop recommendation 24
3.6.2.1 State chart diagram of Navie bayes algorithm used to find success rate 25
3.7.1 Architecture Diagram 26
3.8.1 Context Diagram 26
3.9.1 Use case diagram 27
3.10.1 Sequence Diagram 28
 IOT based crop monitoring and ML based crop recommendation

CHAPTER 1

INTRODUCTION

1.1 ML based Crop Recommendation

India is one of the world's oldest nations that is still engaged in agricultural production.
However, as a result of globalization, the agricultural industry has seen a significant shift in
recent years. A variety of causes have had an impact on the state of agriculture in India. In
order to aid individuals in getting better, several innovative technologies have been created.
One instance of such a strategy is precision agriculture. India is seeing the emergence of
precision agriculture. The practice of gardening "site-specifically" is known as precision
agriculture. In addition to improving our capacity to make better agricultural decisions, it
has also provided us the benefit of more efficient input and output. Precision agriculture has
made important advancements, yet there are still many problems that need to be solved.
There are a variety of methods available that recommend the best inputs for a certain
agricultural area. Crops, fertilizers, and even agricultural methods are suggested by the
systems.
Precision agriculture encompasses a wide range of activities, including crop
recommendation. It is necessary to take into consideration many factors while
recommending crops. These features may be found using precision agriculture in a site-
specific manner, which can subsequently be applied to solve crop selection issues. Despite
the fact that the "site-specific" method has improved the results, it is still important to keep
an eye on how these systems are doing. However, the outcomes of precision farming
approaches might not always be exact. However, in agriculture, it is critical that the
suggestions provided be exact and precise since, in the event of a mistake, it may result in
significant material and financial losses. There are many research projects being carried out
in attempt to develop a crop forecast model that is both accurate and efficient. The article
discusses the needs and preparation that must be done in order to build a software model for
precision farming. It goes into great detail on the fundamentals of precision farming. This
research examines the requirements and planning required for the development of a
software model for precision agriculture. In this chapter, you will learn the basics of
precision farming. Specifically in precision agriculture, the research evaluates classification
Dept of ECE,GMIT,Davangere 2022-23
1
 IOT based crop monitoring and ML based crop recommendation
algorithms and their ability to forecast crop yields. It serves as a model for how Precision
Agriculture (PA) ideas might be applied to modest-sized agricultural businesses. Farmers
may be able to exert some control over the erratic character of their harvests by using open
farms at the level of the individual farmer and crop. One of the techniques used in this sort
of study is assembling, which is discussed below. As a result of this research, one of the
several machine learning techniques now being used in this field is a system that builds an
accurate and efficient model using the voting approach.

Machine learning is a service of artificial intelligence (AI) that offers systems to without
human intervention learn and get better from experience(knowledge) without seeming to be
pre-programmed. Machine learning is concerned with the development of computer
programmers that have the authority to utilize statistical data and to use it for research or to be
taught for their own purposes.

1.1.1 Machine Learning and Classification

Broadly, there are three types of Machine Learning Algorithms.

 Supervised Learning the way it works: this set of A rule is made up of a target / end
outcomes variable (also known as a based variable) that is anticipated to be produced by a
given collection of predictors (impartial variables).
 Unsupervised Learning.
 Reinforcement Learning.

Grouping is a utility of information mining that allots questions that are grouped together
according to board categories or classes The goal of categorization is to achieve a certain
level of efficiency is to accurately predict the board class for every case in the data. For
example,
a classification model may possibly be used to classify loan applicants as near to the
ground, intermediate, or far above the ground credit risks.

 There are observational training sets accessible for supervised machine learning. A
classifier is a programme that executes classification, often in the form of a visible
achievement. Unsupervised machine learning uses a large quantity of data that is sent to
the computer, which then searches for patterns and connections between them. A
machine learning approach called unsupervised learning may be used to find hidden
patterns in the data.

Dept of ECE,GMIT,Davangere 2022-23

2
 IOT based crop monitoring and ML based crop recommendation
 Machine learning combines statistics and computer science to increase the predictive
accuracy. Data scientists and analysts employ machine learning as well as raw data to
forecast or discover trends in the data. In order to obtain the precise crop forecast, there
will be a huge quantity of data in agriculture, and this data is growing every day.

1.1.2 Precision Agriculture

Precision farming means it’s an administration technique that increases efficiency and financial
comes back with a reduced impact on the environment Precision farming is centred on the
application of data innovation to the depiction of inconsistency in the field, and it is becoming
more popular. variable-rate tasks and the basic leadership framework. Precision farming
development includes three technology levels and three strategies. It utilizes a constant
spectrophotometer and was created to depict soil fluctuation in ranchers' fields.

Precision farming, which employs a systems approach, offers a new way to strike a
balance between production and environmental concerns. Precision farming is based on
propelled data innovation. Coordinating rural practices to meet site-particular prerequisites,
depicting and displaying variety in soils and plant species are additionally incorporated into
precision farming. The primary point of accuracy cultivating is to increment monetary returns
and at decreasing the vitality input and the natural effect of farming.

1.2 IOT based Crop monitoring

At the Present days, the farmers chosen most of the methods manually, Irrigation is one of them.
The land to be irrigate time to time. Sometimes this process absorbs more water. Automatic
irrigation system observes the soil moisture content in the soil. It helps to irrigate the plants.
Irrigation of plants takes more time and it has to be done in a justifiable time; It have need large
amount of human resources. All the steps were carried out effectively by humans manually.

Nowadays, some systems use technology to decrease the time taken to water the plants
and it also reduces the number of labors in the field. Water is the main resource which is used
enormously. Mass irrigation is the method used to irrigate the plant. This method represents
huge loss of water, as the water given exceeds the plant’s needs. The excess of water and labor
is becoming more expensive.

The combination of existing methods with Internet of Things and wireless sensor
Dept of ECE,GMIT,Davangere 2022-23
3
 IOT based crop monitoring and ML based crop recommendation
network can develop the agriculture methods efficiently. Concept Internet of Things reads the
data from the sensor and these data are analyzed and displayed to the users. The main aim of
this project is to monitor the agriculture field from any part of world using the concept IOT
without human intervention.

1.2.2 Existing and Proposed System

In conventional irrigation system water is being supplied to field just observing the surface of
the land, there is no mechanism to analyses the moisture content. Motor which supplies water to
field has to be controlled manually.
1.2.2.1 Limitations of existing system
 Presence of farmer all the time is must.
 Location is a barrier to control motors.
 Remote access is not possible.
 Water gets wasted.
 Unplanned water supply leads to loss of crops.

The proposed application works on the principle of IOT, data read from sensor are processed by
the processor (Arduino) then uploaded to database, these data are analyzed and displayed to
users, and user could fetch this information over phone and control and motor using app.

1.2.2.2 Advantages of proposed system

 Farmers could monitor the moisture level from remote place their presence is
not must
 Since it is based on IOT farmer could control motor from far place.
 It does support remote access.
 Water could be used effectively as it is monitored by sensors and uploaded
for farmers notice.
 High yield can be obtained by managing water scientifically.

1.2.2 Feasibility Study

Feasibility study is a study of project where it will verify whether the projected project is
technically and economically possible or not. In feasibility study we analyses the` project to

Dept of ECE,GMIT,Davangere 2022-23

4
 IOT based crop monitoring and ML based crop recommendation
determine the ability to complete the project successfully with all the feasible cost and technical
services. Feasibility study main goal is analyzing the entire feasible test for economical and
operational and function feasibility so that the proposed project cost will be not much
expensive.
Feasibility study is a decision making of project documentation where it will provide
idea to perform the task by using this feasibility report. It also provides the economical and all
the technical cost and resources we needed to perform our task to propose our project.
Feasibility study also used for identifying the scope of our project.
Feasibility study is over all examination of project strength and weakness of project and
required cost to develop our project and also it will provide the problems available in the
offered system and what features we need to include in the existing system to overcome from
the problem of existing system. The feasibility study as follows.
 Technical Feasibility.
 Operation Feasibility.

 Economic Feasibility.
1.2.2.1 Technical Feasibility
The technology feasibility to the proposed system could be summarized as below.
 Data storage
All information is stored in the MySQL database which best suited for distributed application, it
does not require any additional drivers to make transactions with database it is added advantage,
the proposed system works on the principle of client server architecture it does not require
additional software at client side to make the system functional.
 Web server
Server which needs to execute continuously and serve client requests reliably, in this project
Apache web server is used for this purpose which provides service at 3306 dedicated port,
server has the potential to work concurrently and provide service to clients instantly, it works
best on Linux platform as it is a multiuser operating system. This server executes the PHP
programs efficiently.
 Messaging
In the proposed system SMS are sent via GMS modem, AT (Attention Commands) commands
are used to communicate with the modem each command is an alphanumeric code with
predefined meaning and commands takes parameters.

Dept of ECE,GMIT,Davangere 2022-23

5
 IOT based crop monitoring and ML based crop recommendation
 Server Scripting
In the proposed system server-side scripting is done using PHP language, the language supports
both procedural and object-oriented approach, it is flexible language as it supports variant data
type, for each loop with key value concept which is very help full to navigate through arrays of
different types.
1.2.2.2 Operational feasibility
The planned system is an effort to make use of web applications and internet services to design
a well generalized web site which act as communication media between users and the order
system and law.
Since users have rights to make complaint against the crime after being verified as a
authenticated users. only the register users can make a complaint over online regarding crime
and miss guidance and even for providing illegal information on web sites which may create
problems to the accessed users.
One more benefit of using web sites is that any complaint regarding crime and illegal
work complaints should be made online only by using provided facilities on web sites which
avoids the waste of time. There is no need to visit particular stations frequently to complaint
against crime. Complaints given by users will be handled by the authorized officials and they
provide solution to their solutions to their problems.
Operational feasibility monitors all these progresses to provide good service to the
public/users so they can access the useful information through web sites and design their own
applications and users are allowed to download the useful information they can upload their
works and ideas on web sites they can also give feedback to the accessed websites
1.2.2.3 Economic Feasibility

Economic practicableness determines whether or not the specified code is capable of generate
currency gain for a company. It involves the value of code development, calculable price of
hardware and code, price of maintenance. The aspects mainly verify whether the development is
cost-effective. In our proposed system, the development cost is optimized. The maintenance
cost of the proposed system will be less.

Dept of ECE,GMIT,Davangere 2022-23

6
 IOT based crop monitoring and ML based crop recommendation

1.3 Overview
Machine Learning (ML) based crop recommendation and IoT (Internet of Things) based crop
monitoring represent advanced technologies revolutionizing agriculture. ML-based crop
recommendation systems analyze vast datasets, including soil properties, weather patterns, historical
crop performance, and even economic trends. They use algorithms to predict the most suitable crops
for a given area, optimizing for factors like yield, disease resistance, and market demand.
These systems can consider local environmental conditions, previous crop rotations, and pest
infestations, making recommendations highly tailored to specific locations. This approach helps
farmers avoid trial-and-error methods, leading to more efficient farming practices, reduced waste,
and potentially higher profits.
On the IoT side, sensors deployed in the field collect real-time data on soil moisture,
temperature, humidity, and nutrient levels. This data is transmitted to cloud-based platforms, where
it's analyzed to monitor crop health and growth. Farmers receive updates on their crops' conditions,
allowing for timely interventions like watering, fertilizing, or pest control.
IoT devices can also be equipped with cameras and drones for visual inspection, giving
farmers a comprehensive view of their fields. This technology helps in early detection of diseases or
deficiencies, enabling prompt action to mitigate risks.
Integration of ML and IoT in agriculture leads to a more data-driven approach. While ML
offers insights for long-term planning and crop selection, IoT provides the real-time monitoring
necessary for day-to-day management. This combination enhances precision agriculture, leading to
increased crop yields, reduced environmental impact, and more sustainable farming practices.
Farmers using these technologies can adapt more quickly to changing conditions, such as
climate change or market shifts. The data collected also helps in researching new agricultural
methods and crop varieties, further advancing agricultural science.
In summary, ML-based crop recommendation and IoT-based crop monitoring represent a
significant step forward in agricultural technology. They provide farmers with powerful tools for
improving efficiency and productivity, while also contributing to the broader goals of sustainable
and resilient agriculture.
1.4 Problem Statement

Mining information can be utilized in agriculture to improve agricultural practices and decision
making by analysing an array of abiotic and biotic factors. In India’s economy, agriculture has a
significant role and in finding the jobs. Lack of knowledge about which crops would be best suited

Dept of ECE,GMIT,Davangere 2022-23

7
 IOT based crop monitoring and ML based crop recommendation
for a specific piece of land is the most significant obstacles that farmers must overcome. This issue
has the possibilities to have enormous repercussions for the productivity, profitability, and long-term
viability of agriculture. Farmers are able to arrive at decisions on crop selection that are more precise
and appropriate when they embrace practices of precision agriculture while taking into consideration
the particular qualities of their land. This can guide to increased yields, decreased costs of inputs, and
an overall improvement in the agricultural sector’s ability to remain sustainable.
1.5 Objectives
 To implement machine-learning algorithms for appropriate crop prediction for specific areas.
 To predict authentic results using trained model (dataset).
 To help the farmers in choosing right crop to grow.
 To reduce crop loss burden for farmers to improve their resilience to risks, and enhance
their overall agricultural productivity and income.

1.6 Motivation of the project

The motivation for developing an ML-based crop recommendation and IoT-based crop monitoring
system stems from the critical need to address challenges in modern agriculture. Traditional farming
practices often face uncertainties related to climate change, resource optimization, and market
demands. Here are key motivations for such a project:

 Optimizing Resource Utilization: Farmers often encounter challenges in determining the

most suitable crops for their specific conditions. ML-based crop recommendation can provide
personalized suggestions based on historical data, soil analysis, and climate patterns, enabling
optimal use of resources like water and fertilizers.
 Increasing Crop Yields: ML algorithms can analyse vast datasets to identify crop varieties
that thrive in particular environments. This targeted approach can lead to increased yields,
contributing to food security and economic stability for farmers.
 Precision Agriculture for Sustainability: Integrating IoT sensors for real-time crop
monitoring allows farmers to adopt precision agriculture. This leads to more sustainable
practices by minimizing overuse of resources, reducing environmental impact, and mitigating
the risk of crop diseases.
 Climate Resilience algorithms can adapt to changing climate conditions, providing farmers
with insights on crops resilient to environmental shifts. This proactive approach helps farmers
Dept of ECE,GMIT,Davangere 2022-23
8
 IOT based crop monitoring and ML based crop recommendation
anticipate and mitigate the impacts of climate change on their agricultural practices.
 Data-Driven Decision Making: The combination of ML and IoT enables farmers to make
informed decisions based on real-time and historical data. This empowers them to respond
promptly to crop health issues, market trends, and environmental changes, enhancing overall
efficiency.
 Economic Empowerment: By optimizing crop selection and improving yields, the project
aims to contribute to the economic well-being of farmers. Enhanced productivity and
resource management can lead to increased income and financial stability for agricultural
communities.
 Technological Innovation in Agriculture: Embracing ML and IoT technologies in
agriculture represents a paradigm shift towards innovation. The project serves as a catalyst
for the adoption of advanced technologies in the farming sector, fostering agricultural
modernization.
 Global Food Security: With a growing global population, ensuring food security is
paramount. ML-based crop recommendation and IoT-based monitoring can play a crucial
role in meeting the rising demand for food production by making agriculture more efficient
and resilient.
In conclusion, the motivation behind developing an ML-based crop recommendation and IoT-
based crop monitoring system is rooted in addressing the pressing challenges faced by modern
agriculture. The project aims to empower farmers with advanced technologies, promoting
sustainability, resilience, and economic prosperity in the agricultural sector.

Dept of ECE,GMIT,Davangere 2022-23

9
 IOT based Crop monitoring and ML based crop recommendation

Chapter 2

LITERATURE REVIEW

1] S. Pudumalar, E. Ramanujam, R. HarineRajashree, C. Kavya, T. Kiruthika, J. Nisha

depicts the Yield Recommendation System for Precision Agriculture.
This author introduces the fundamentals of precision farming before moving on to developing a
model. On the rancher and product levels, it shows the accuracy of horticulture needs for small, open
ranches. In India, the most common issue that farmers encounter is that they do not pick the most
suitable crop for their particular soil type. As a consequence of this, they will see a substantial
decrease in their output. We want to reach wish to do this by using the most cutting-edge
technologies, such as SMS and email, as the primary destinations for advice warning
administrations, in order to reach even the tiniest rancher at the level of his or her smallest harvest
plot. This model was created expressly for the state of Kerala, where typical holding sizes are much
less than those of the rest of India.
2] Saiyyad Mohmmad Ali, Muzffar Ali depicts Study of order calculations for detailing yield
forecast exactness in accuracy
Agriculture planning is essential in every nation, and it is especially significant in developing
countries. The agriculture sector generates a variety of products, including food, raw materials for
industry, economic stimulus, and job creation. There is a wealth of information available on the
variables that influence the input and output of the agricultural industry. As technology progresses,
new data mining methods are being developed and implemented. These data mining methods may be
used to examine the multidimensional, time specific data of the agricultural sector in order to get
useful information from it that can be used to help the economy grow. Throughout this paper, we
will examine various approaches taken by different researchers to deal with the analysis of the
agriculture sector in order to provide an effective way to increase production while also making the
most of available resources, which in turn will increase the economy in this sector.
3] Supriya D M Data Mining Approach for Soil Behavior Analysis and Crop Yield Prediction
Yield prediction contributes to the proper selection of crops for sowing which is very popular
among farmers these days [4]. An interesting challenge is predicting and yielding of the crops. In
earlier days yield prediction was performed by considering the farmers experience on a particular
field and crop. Yield prediction uses the data mining techniques 4 in 4 order 2 to 9 predict the
category of the analyzed soil datasets. Yielding of crops will be indicated by prediction Naïve
Dept of ECE,GMIT,Davangere 2022-23
10
 IOT based Crop monitoring and ML based crop recommendation
Bayes and K-Nearest

Dept of ECE,GMIT,Davangere 2022-23

11
 IOT based Crop monitoring and ML based crop recommendation
Neighbor methods are used for predicting the crop yield.
4] E. Manjula, S. Djodilta Choum depicts An Agricultural Yield Prediction Model.
In prior days yield expectation was performed by considering agriculturists encounter on specific
field and product [5]. One of the significant issues that remaining parts to be fathomed in light of
accessible information is yield forecast. One of the better decisions for this is information mining
systems. For assessing what's to come year's harvest generation distinctive Data Mining systems are
utilized and assessed in farming. A framework to anticipate trim yield from past information
proposed and executed in this exploration. It can be accomplished by applying affiliation administer
mining on farming information. Formation of a forecast demonstrate which might be utilized to
future expectation of harvest yield is focused in this exploration. In this paper it shows a short
examination of product yield expectation utilizing information mining strategy in view of affiliation
rules for the chose locale i.e., locale of Tamil Nadu in India.
5] Anshal4Savla, Himtanaya*Bhadada, ParulDhawan, Vatsa Joshi depicts the use of machine
learning techniques for yield prediction on delineated zones in precision agriculture.
Normalization is the process through which information in a database is sorted and organized. There
are a lot of stages involved. A large quantity of data is gathered in agriculture, and a variety of data
mining methods are used to make effective use of this information. It is possible to diffract a table
into a lot of smaller tables that are less repetitious. Outside keys are defined in old tables that refer to
the important keys in new tables, and no data is lost in the course of the process. The information
standardization process has been completed in order to disconnect it. In order for the progressions
made in the characteristics of one table to be communicated to the whole database as efficiently as
possible. From the perspective of data mining, this boils down to a variation of spatial clustering that
is constrained by the requirement that the resultant clusters be geographically largely continuous in
nature. The ultimate aim is to guarantee that the information generation of tables and the
characterization of relationships between them is completed while keeping this in mind. Data
normalization is a helpful idea in order to arrange a large data collection of information. The data is
divided into intelligent gathers, with each gathering correctly depicting the little bit of information
that was originally collected. It is possible to modify the information in a large database efficiently.
By adjusting the date range to be more consistent It has been accomplished to make information
more accessible and to manage it more quickly.

6] Suma, N., Samson, S. R., Saranya, S., Shanmugapriya, G., & Subhashri, R. (2017)
IOT based smart agriculture monitoring system. International Journal on recent and Innovation

Dept of ECE,GMIT,Davangere 2022-23

12
 IOT based Crop monitoring and ML based crop recommendation
Trends in computing and communication, 5(2), 177-181.
7] Prathibha, S. R., Hongal, A., & Jyothi, M. P. (2017, March)
IOT Based monitoring system in smart agriculture. In 2017 International Conference on Recent
Advances in Electronics and Communication Technology (ICRAECT) (pp. 81-84). IEEE. G.
8] Ramkumar, P. Parkavi, K. Ramya and M. S. Priya
"A Survey on Sar Images Using Image Processing Techniques,” 2020 6th International Conference
on Advanced Computing and Communication Systems (ICACCS), 2020, pp. 1097-1100, doi:
10.1109/ICACCS48705.2020.9074261.
9] Chandoul Marwa, Soufiene Ben Othman and Hedi Sakli
"IoT Based Low-cost Weather Station and Monitoring System for Smart Agriculture", International
Conference on Sciences and Techniques of Automatic Control and Computer Engineering, 2020.
Mohamed Rawidean Mohd Kassim, “IoT Applications in Smart Agriculture: Issues and Challenges",
IEEE Conference on Open Systems, 2020 G.
10] Ramkumar, A. G, S. K. M, M. Ayyadurai and S. C
"An Effectual Underwater Image Enhancement using Deep Learning Algorithm," 2021 5th
International Conference on Intelligent Computing and Control Systems (ICICCS), 2021, pp. 1507-
1511, doi: 10.1109/ICICCS51141.2021.9432116

Dept of ECE,GMIT,Davangere 2022-23

13
 IOT based crop monitoring and ML based crop recommendation

Chapter 3
METHODOLOGY

When it comes to software development, high-level design (HLD) is used to describe the architecture
that will be utilized. The architecture diagram offers a high-level perspective of an entire system,
indicating the major components that would be created for the product as well as their interactions. It
is used to design a product's architecture. When describing the HLD, non-technical to slightly
technical words are used, which should be comprehensible to the system's administrators. Low level
design, on the other hand, exposes programmers to a greater extent the logical precise design of each
of these components.

The design at the highest level is called high level design which is used to design the software
related requirements. In this chapter complete system design is generated and shows how the
modules, sub modules and the flow of the data between them are done and integrated. It is very
simple phase that shows the implementation process. The errors done here will be modified in the
coming processes.

3.1 Design Consideration

 User Characteristics/Roles

The user characteristics are as follows:

The end users of the application are framers, they just need to select the location of the filed by
means of maps, using the REST API of soil grids the parameters of the soil are fetched, analyzed and
crop is recommended.

 Input Requirements
 Farmers needs to feed the geo location by means of latitude and longitude, it is done
using android app, location will be fetched automatically
 Using REST full API of soil grids, various features of the soil are fetched
 Collected details are analysed with trained dataset
 Crop Recommendation is done.
 Output Requirements

Information about recommended crops is displayed.

Dept of ECE,GMIT,Davangere 2022-23

14
 IOT based crop monitoring and ML based crop recommendation

3.2 The Organization of the System

It is the conceptual model of a system that specifies its structure, behavior, and many other aspects of
its operation, known as its system architecture. When describing a system, an architectural
description organizes the information so that it can be reasoned about the various components of the
system.

Dataset
Collection
Feature Extraction
Training data

Testing Data

ENSEMBLE MODEL
Recommendation System K-Nearest Neighbour
Naïve Bayes
Rules Induction

Crop to be yield

Figure 3.2.1: System Architecture of the proposed system.

The following guidelines were taken into account while designing the application shown in
figure 3.2.1:
 A3 design should exhibit hierarchical organization that makes the intelligent use of control
among components of the software.

Dept of ECE,GMIT,Davangere 2022-23

15
 IOT based crop monitoring and ML based crop recommendation

 A design should be modular i.e. the software should be logically partitioned into components
that perform specific function and sub function.
 A design should lead no interfaces that reduce the complexity of connections between and
modules with the external environment.

The proposed system has the following steps for recommendation of crop.

1. Dataset Collection
2. Training Data
3. Testing Data
4. Feature Extraction
5. Ensemble Model
6. Recommendation System

3.2.1 Dataset Collection

A data set is a collection of information or a collection of data. When presented in a tabular

format, the information in this collection is usually more digestible. Every column contains
information about a certain variable. And each row refers to a certain member of the data
collection, which is in accordance with the query. This is a subset of data administration.
Measurement error minimization is only possible if the prerequisites for success are met in the
data collection and capture setups. When it comes to measurement error, the data collection and
capture procedures may be significant contributors. Measurement error can be described as the
difference between a variable's value given by the responder and the actual, but unknown, value
of that variable.

3.2.2 Training Data

The training data is a collection of data that is used to assist a programme understand how to use
technologies such as neural networks to learn and generate advanced results. It is important to
note that the training data is not the same as the test data. Additional data sets, referred to as
validation and testing sets, may be used to supplement the first data set. The trained data is used
to ensure that the machine recognises patterns in the data, the cross-validation data is used to
ensure that the algorithm used to train the machine is more accurate and efficient, and the test data
is used to determine how well the machine can predict new answers based on its training.

Dept of ECE,GMIT,Davangere 2022-23

16
 IOT based crop monitoring and ML based crop recommendation

Agribusiness data has been gathered from the agricultural agency and is presented here. This is a
large amount of information that include the soil characteristics as well as the crop table. In this
database are the data that are based on real farmers, the crops that are grown by them, and the
results of their labour.

3.2.3 Testing Data

Unrelated to the training dataset, a test dataset is a collection of data that follows the same
probability distribution as the training dataset, despite the fact that it was generated independently
of it. If a model that was fit to the training dataset also fits the test dataset well, it is likely that
only little overfitting has occurred. In most cases, a greater fit of the training dataset compared to
the test dataset indicates overfitting of the dataset.

With the soil grid website, we provide the location co-ordinates, which are really the
latitude and longitude of that specific place, as input. The soil grid website then extracts the soil
characteristics of that particular location with the use of the android application. Initially, this data
is utilised as testing data, and then it is used for the actual processing.

3.2.4 Extraction of Characteristics

With the use of feature extraction, dimensionality reduction may be accomplished by dividing an
initial collection of raw data into more manageable groupings that can then be processed. The high
number of variables in these big data sets necessitates the use of a significant amount of computer
resources in order to process them. When you need to minimize the number of resources required for
processing without losing any essential or relevant information, the feature extraction method may
be quite helpful. Feature extraction may also help to minimize the quantity of duplicated data that is
generated for a particular study.

In this section, we extract soil characteristics such as pH, carbon, slit, sand, and clay from the
training dataset.....These characteristics are utilised in the computations portion of the process,
which
also aids in the recommendation system and ensemble model development.

3.2.5 Ensemble Model

According to machine learning terminology, classification is considered an instance of supervised

learning, which is learning that takes place in the presence of a training set containing properly
Dept of ECE,GMIT,Davangere 2022-23
17
 IOT based crop monitoring and ML based crop recommendation
recognised observations (observations with the right labels). Clustering is the unsupervised

Dept of ECE,GMIT,Davangere 2022-23

18
 IOT based crop monitoring and ML based crop recommendation

method that corresponds to this supervised procedure, and it entails categorizing data into groups
based on some measure of intrinsic similarity or distance.

3.2.6 Recommendation System

This is the system where the manipulated are resulted. Here based on the result of KNN algorithm
progress the crop is recommended. According to the value of k different crops are recommended and
then over the voting system best crop is recommended. Recommendation system can recommend
more than one crop.

3.3 Specification using Use Case Diagram

Essentially, a use case diagram, as shown in figure 3.2, is a visual depiction of an individual's
engagement with a system that depicts their connection with the many use cases in which they are
engaged.

Launch app

Allow
location

Click
recommend

algorithm Users
Get crop
name

Figure 3.3.1: The System's Use Case Diagram is shown above.

3.4 Module Specification:

Module Specification is the way to improve the structure design by breaking down the system into

Dept of ECE,GMIT,Davangere 2022-23

19
 IOT based crop monitoring and ML based crop recommendation

modules and solving it as independent task. By doing so the complexity is reduced and the modules
can be tested independently.

 Ensemble Module

 Name of the Module: K Nearest Neighbor

 Actor: System
 Use Cases: Fetch test data, Find Euclidian, Sort records based on K, Create voting
 Functionality: The primary purpose of this module is to provide recommendations on
which crop to plant.
 Description:
Figure 3.3 shows the use case diagram of classification using KNN module. In this use case diagram,
there are four use cases and one actor. In the first use case, the system fetches test data. In second use
case the Euclidian distance is obtained. In the third use case the records are sorted based on k. In the
fourth use case the voting is done and the favorable or the recommended crop which is best suited is
suggested here.

Fetch test
data

Find Euclidian distance

Sort records based on K

System
Create voting table

Figure 3.4.1: Use Case Diagram of Ensemble using KNNmodule

Dept of ECE,GMIT,Davangere 2022-23

20
 IOT based crop monitoring and ML based crop recommendation

 Name of the Module: Naïve Bayes

 Actor: System
 Use Cases: Use test data parameters, create frequency table, find probability of each test
parameter, find final probability using all test parameters
 Functionality: The primary purpose of this module is to provide recommendations on which
crop to plant.
 Description: Figure 3.4 shows the use case diagram of classification using Naive bayes
module. In this use case diagram, there are four use cases and one actor. In the first use case,
the system takes preprocessed test data parameters. In second use case the classifier creates
frequency table. In the third use case classifier finds the probability of each test parameters.
In the fourth use case final probability is computed using all the test parameters.

Use test data parameters

Create Frequency
Table

Find probability of each test parameter

Naïve base algorithm

Find final probability using all test parameters

Figure 3.4.2: Use Case Diagram of Ensemble using Naïve Bayes Module.

Dept of ECE,GMIT,Davangere 2022-23

21
 IOT based crop monitoring and ML based crop recommendation

3.5 Data Flow Diagram

In accordance with the meaning of the terms, it is the process itself that is described in depth, such as
how data is transferred between the various processes. Data flow diagrams (also known as DFDs) are
used to visually depict the movement of data in a corporate information system. They are also known
as data flow charts. In a system, DFD defines the procedures that are used to move data from the
input to the file storage and to the production of reports. It is possible to split data flow diagrams into
two types: logical and physical. The flow diagram shown in the following picture 3.5 is comprised of
three parts: the input, the process, and the output.
After each procedure, the data that is transferred between the systems must be defined,
which is done via the use of a diagram known as a data flow diagram. Most of the time, it is the first
stage in the process of developing any of the systems that will be put into place. It also indicates
where the data began, where it flowed, and where it was eventually kept. The input picture is
preprocessed, following which feature extraction is performed, followed by classification once the
model has been trained, and finally illness prediction is performed at the last phase.

Figure 3.5.1: Data flow Diagram of System.

Dept of ECE,GMIT,Davangere 2022-23

22
 IOT based crop monitoring and ML based crop recommendation

As shown in figure 3.5, the data flow diagram illustrates the manner in which data is handled by a
system in terms of inputs and outputs. It is concerned with the movement of information. Data stores
are the storage locations for information in a system. These are the0pipelines0through which packets
of information are transported from point A to point B. Label is a term used to describe arrows that
have the name of the data that is moving through them.

3.5.1 Data Flow of KNN Module

In machine learning, the K-nearest neighbors (KNN) method is an example of a supervised learning
technique that may be used to both classification and regression prediction problems. However, in
industry, it is mostly utilized to solve classification and prediction issues. The following two
characteristics would accurately characterize KNN. It is referred to as a lazy learning algorithm since
it does not have a dedicated training phase and instead utilizes all of the data for training and
classification at the same time. KNN is sometimes referred to as a non-parametric learning algorithm
since it makes no assumptions about the data it is learning from.

When the user enters the location, a HTTP request is sent to the apache server and a model is
created. Information regarding the location is acquired from rep table. Referring model is created and
thus a matrix table is generated. This creates a voting table referring to matrix table as shown in
figure 3.6

Figure
3.5.1.1 : Data Flow Diagram of
KNN.

Dept of ECE,GMIT,Davangere 2022-23

23
 IOT based crop monitoring and ML based crop recommendation

3.5.2 Modeling the Data Flow of the Nave Bayes Module

Bayes' theorem has a wide range of applications in probability theory and statistics. The probability
that you have never heard of this theorem in your life is very low. In fact, the Naive Bayes Data
Flow Diagram example includes five processes and four data storage, as shown in the diagram.
Based on the graphic, we may deduce that a user has sent a http request. Get information receives the
request, forwards it to process information, stores it in the Freq table data store, and updates the
process information details in the Freq table data store, forwards it to Referring freq table and
generating prob table, stores it in the total table data store, forwards it to the Generating final result
process, stores it in the total result table data store, and forwards it to the Referring freq table and
generating prob table process, stores it in the total result table data store In addition, the procedure
provides the user with a suggested crop. Because of this theorem's use in the field of machine
learning, one of the most highly decorated algorithms has been created. In this post, we will study all
there is to know about the Naive Bayes Algorithm, as well as its variants for various applications in
machine learning.

Figure 3.5.2.1: Data Flow Diagram of Naïve Bayes.

3.6 Diagram of the State Chart
The state chart diagram is one of five diagrams in the Unified Modeling Language (UML) that are
used to describe the dynamic nature of a system. They describe several states of an item during its
lifespan, and these states are altered as a result of various events. State chart diagrams are helpful for
modelling reactive systems because they are easy to understand. Reactive systems may be described
Dept of ECE,GMIT,Davangere 2022-23
24
 IOT based crop monitoring and ML based crop recommendation

as systems that react to external or internal stimuli in a predictable manner. A flowchart is a visual
representation of the sequence of actions and decisions that are anticipated to be made throughout the
execution of a process. Each step forward in the grouping has been recorded. Every progression is
linked by interface lines and directional bolts, which are used to connect each step. It is one of the
most effective business tools available. As shown in figure 3.8, it transports the means in a process
effectively and efficiently, while adhering to proper outline and growth.

Input location usi Location sent Fetch the soil parameter

ng
from soil grid
GPS

Soil parameter
Soil parameter sent sent

Find the Find the high yield

recommendation crop using KNN percentage using Naïve Bayes

Resultant crop yield

Resultant crop

Figure 3.6.1: State Chart Diagram of System.

3.6.1 State Chart Diagram of KNN

The below diagram shows various processing steps being carried out while recommending a crop
using KNN algorithm.

When user launches the app using an android phone the current location of the user is fetched and
this location info is passed to soil grid API to fetch soil parameters of the current location, these
values are treated as test data. When user clicks on the recommend button KNN is implemented to
recommend a crop.

Dept of ECE,GMIT,Davangere 2022-23

25
 IOT based crop monitoring and ML based crop recommendation

The first step is building the model having Euclidian distance (ED), in the next step the records
having least distance are selected and the number of records depends on the value of K (Neighbors).
These neighbor records are grouped to create voting table and based on the number of votes
(Maximum) received crop is recommended as shown in figure 3.9.

Fetch ph, carbon, Data sent

Build model to find ED
sand, slit, clay and click recommend

ED selection
Matching data Processing data

Select k nearest Generate voting table

neighbours and recommend crop

Recommended crops Selection of crop

Figure 3.6.1.1: State Chart Diagram of KNN Algorithm used for Crop Recommendation.

3.6.2 State Chart Diagram of Naïve Bayes

Above below shows the various steps being fallowed to apply naïve base algorithm. Soil parameters
of test data are used to find the probability, in the first step it is required to build a frequency table,
this table includes count of each parameter like number of records having value yes for particular ph.
and number of parameters having no for each ph. value. Similarly count of records based on yes or
no value of each parameter is found.

Having found the counting of each parameter probability has to be obtained using the formula
number yes parameters divided by total number of yes present in training data, same rules applied for
no parameter too.

Having found the probability, it is required to fetch the nearest value of record matching our test data
and this step should be done for each soil parameter and records should be stored in a table for later
reference.
Dept of ECE,GMIT,Davangere 2022-23
26
 IOT based crop monitoring and ML based crop recommendation

Now find out the avg probability using the table containing probability of each parameters shown in
figure

Consider the soil Data sent Build Frequency Table

parameter of the test data

Nearest value
Selection of data Processing data

Get the avg probability

Find probability of using the formula
each parameter

Average of
Probability of data probability

Figure 3.6.2.1: State Chart Diagram of Naïve Bayes Algorithm used to find Success Rate.

3.7. Crop monitoring

The main aim of the system perspective is to reduce the complexity of the system. Here the system is
described as a whole not as an isolated individual system. Hence the relationship between the system
and the environment is to be considered. This system perspective study also gives information
regarding its behavior and properties; this may also include the interactions that the proposed system
does with the present environment. This part of the report gives information about the system that is
proposed and its relationship with the surrounding environment. The main components of this
mobile application are as follows:

Dept of ECE,GMIT,Davangere 2022-23

27
 IOT based crop monitoring and ML based crop recommendation

3.7.1. Architecture diagram

3.8 CONTEXT DIAGRAM

Fig 3.8.1 Context diagram

Description: The Fig 3.8.1 describes the end user data which process to the base station and
sensor network for the data storage through the internet as a medium and which indicates the data

Dept of ECE,GMIT,Davangere 2022-23

28
 IOT based crop monitoring and ML based crop recommendation

fetch by the sensors to the perspective’s users.

3.9 Use case Diagram

Get notifications

Launch app

View status

Farmer
Control the motor

Figure 3.9.1 Use Case Diagram

 Description: A use case diagram at its simplest is a representation of a user’s interaction

with the system that shows the relationship between the users and the different use cases in
which the user is involved.

Dept of ECE,GMIT,Davangere 2022-23

29
 IOT based crop monitoring and ML based crop recommendation

3.10 Sequence diagram

: Sensor : Centralized Server Farmer

2. Upload to server
1. Collect 3. Display information to farmer
moisture
content from
soil

4. Request to Control the device

5. Request is processed and motor is either turned on or off

Figure 3.10.1 sequence diagram

 Description: The Figure 5.2 describes the sequence flow of data which from senor to
centralized server and also to the farmer network for the data storage through the internet as a
medium.

Dept of ECE,GMIT,Davangere 2022-23

30
 IOT based crop monitoring and ML based crop recommendation

3.11 Dataflow diagram

1. Fetch moisture from
soil

2. Check the threshold values &turn on/off LED’s

ON/OFF
RED, Green LED

3. Establish connection with server using Wi-

Fi

4. Upload to database using PHP script

Moist table

5. Control motor based

on user response ON/OFF
Motor

Figure 3.11.1 Dataflow diagram

 Description: The Figure 5.3 describes the data fetch from the moisture content of the soil
to user data which process to the data base and sensor network and control through the motor
and moist table.

Dept of ECE,GMIT,Davangere 2022-23

31
 IOT based crop monitoring and ML based crop recommendation

3.12 Activity Diagram

IOT System

Fetch soil parameters Control devices

Upload to server

Figure3.12.1 Activity diagram

 Description: The Figure 5.4 describes the data fetch from the moisture content of the soil
to user data which process to the data base and sensor network. to fetch soil parameters,
upload server and control devices.
3.13 Database Design
3.13.1 Table used in the project

Humidity table

Column Type

Sino int (11)

Hum int(11)

Uptime Time

This table is used to store the values sent from sensors

Dept of ECE,GMIT,Davangere 2022-23

32
 IOT based crop monitoring and ML based crop recommendation

3.13.2 Sample data

Sino hum Uptime

1 4095 19:26:06

2 4095 19:26:12

3 4095 19:26:18

4 256 19:26:25

5 208 19:26:31

6 192 19:26:37

7 184 19:26:43

8 4095 19:26:49

9 4095 19:26:55

10 209 19:27:01

11 232 19:27:07

12 228 19:27:13

13 224 19:27:22

14 225 19:27:28

15 224 19:27:34

16 240 19:27:40

17 240 19:27:46

18 240 19:27:52

19 248 19:27:58

20 248 19:28:04

21 257 19:28:10

Dept of ECE,GMIT,Davangere 2022-23

33
 IOT based crop monitoring and ML based crop recommendation

Sino hum Uptime

22 272 19:28:16

23 276 19:28:22

24 290 19:28:28

25 304 19:28:35

26 320 19:28:41

27 322 19:28:47

28 336 19:28:53

29 348 19:28:59

30 368 19:29:05

Above table shows the sample data being read from sensor
3.14 Pin configuration of ESP32

Figure 3.14.1 configuration of ESP32IC

Dept of ECE,GMIT,Davangere 2022-23

34
 IOT based crop monitoring and ML based crop recommendation

Above figure shows the pinconfiuration of ESP32 Kit which is used in our project send information
read from sensor to database using Wi-Fi

Dept of ECE,GMIT,Davangere 2022-23

35
 IOT based crop monitoring and ML based crop recommendation

Dept of ECE,GMIT,Davangere 2022-23

36
 Pick and Place Movable Machine

Chapter 4
HARDWARE AND SOFTWARE REQUIREMENTS

4.1 Hardware Requirements

4.1.1 Processor - i3 or any compatible
4.1.2 Hard Disk - 500GB
4.1.3 ESP32 chip
4.1.4 Soil moisture sensor
4.1.5 Water pumps
4.1.6 LED s

4.1.1 Processor - i3 or any compatible

Figure 4.1.1 processor-i3 or any compatible

1Performance hybrid architecture combines two core microarchitectures, Performance-cores (P-
cores) and Efficient-cores (E-cores), on a single processor die first introduced on 12th Gen Intel®
Core™ processors. Select 12th Gen and newer Intel® Core™ processors do not have performance
hybrid architecture, only P-cores or E-cores, and may have the same cache size. See ark.intel.com for
SKU details, including cache size and core frequency.
2Built into the hardware, Intel® Thread Director is provided only in performance hybrid
architecture configurations of 12th Gen or newer Intel® Core™ processors; OS enablement is
required. Available features and functionality vary by OS.

Dept of ECE,GMIT,Davangere 2022-23

37
 Pick and Place Movable Machine

4.1.2 Hard disk

Figure 4.1.2 Hard disk

Hard disk drive, hard disk, hard drive, or fixed disk, is an electro-mechanical data storage
device that stores and retrieves digital data using magnetic storage with one or more rigid
rapidly rotating platters coated with magnetic material

4.1.3 ESP32 Chip

Figure 4.1.3 ESP32 Chip

This above Fig 4.1.3 describes the Arduino board which is main bridge for connecting
with other peripherals.

Dept of ECE,GMIT,Davangere 2022-23

38
 Pick and Place Movable Machine
4.1.4 Soil Moisture Sensor

Figure 4.1.4 Soil Moisture Sensor

This above Figure 4.1.4 describes the Moisture Sensor used to collect the data from the soil.

4.1.5 Water pump

Figure 4.1.5 Water pump

This above Figure 4.1.5 describes the submersible water pump which determines the flow
of water.

Dept of ECE,GMIT,Davangere 2022-23

39
 Pick and Place Movable Machine
LED’s

Figure 6.2.5 LED’s

The above Figure shows the LED’s (Light Emitting Diodes) being used in the project, they
operate at a voltage range of 1.8 to 3.3 vin this project they are used to indicate the moisture
level of soil, red led indicates soil is dry and blue indicates soil is wet.

4.2 Software Components

4.2.1 Front end - HTML, JAVA Script, Boot Strap Frame work
4.2.2 Back end - My SQL
4.2.3 Server Scripting Language- PHP
4.2.4 IDE - Android studio Smart phone running with android OS
4.2.5 Blynk dash board

The procedure of executing system with the target of finding error is outlined as testing. It
can also be defined as the process that defines, isolates, subjects to rectification of defects, and so
that the customer satisfaction is reached at last with the assurance of the system is free from defects.
Software testing is a very important element of the quality assurance and it represents the SRS,
designing, coding and implementation of the system proposed.
4.2.1 Levels of testing:
 Test Planning:
Test plan is the document that gives the information regarding the procedure that is to be followed in
performing various tasting on the whole application.
This document involves scope and objectives of the testing, areas that are to be tested and areas that
should not be tested, scheduling of resources available, the area that need to be automated and
Dept of ECE,GMIT,Davangere 2022-23
40
 Pick and Place Movable Machine
various tools that are used for testing.
 Test Development:
Test development involves development of test cases and their procedural preparation i.e. description of
the developed test cases.

Types of testing:
Various types of testing that are done on the system are as follows:
i. Unit testing
ii. Integration testing
iii. System testing

 Unit testing: As the name itself says, this type of testing is done on small units of the system.
A part of the system is considered as a unit and its testing is done. If as an example, login
page considered; the user or the administrator can enter into their respective home pages only
after giving the valid username and password. This part of validating a system, by
considering Login as a unit can be said as a unit testing.
 Integration testing: This part of testing deals with the testing procedure. It involves, testing
of various integrations of several units. It checks whether the system is functioning correctly
when two or more units are integrated together. This part of testing gives information about
order of arrangements of various units, integrating modules, systems, sub-systems and the
entire system as a whole.
 System testing: This testing technique deals with the process of testing the system as a
whole. At the end of each project, all defects are removed and the interface errors are
uncovered in order to achieve the good functioning of the whole system. This testing
technique can be called as the final part of whole testing process.

Dept of ECE,GMIT,Davangere 2022-23

41
 Pick and Place Movable Machine

4.2.5 Blynk dash board

Figure 4.2.5 Blynk dashboard

The above figure shows the dashboard of blynk displayer at farmers mobile. It has labels to
display temperature humidity and soil moisture level, with button farmer could control the pump.

Dept of ECE,GMIT,Davangere 2022-23

42
 Pick and Place Movable Machine

Chapter
6
RESULTS AND DISCUSSION

Figure 6.1. Snapshot of Location Based Crop Recomnadation

System

Figure 5.1.2 Snapshot of Crop Recomandation

Dept of ECE,GMIT,Davangere 2022-23

43
 Pick and Place Movable Machine

Figure 5.1.3 Result of Crop Recomandation System

Dept of ECE,GMIT,Davangere 2022-23

44
 Pick and Place Movable Machine

Chapter 7

ADVANTAGES, DISADVANTAGES AND APPLICATIONS

7.1 Advantages

 Machine learning based process always produces good result

 Choosing right crop according to soil parameter is the need of the hour hence farmers will

be guided for this purpose

7.2 Disadvantages

 Having devices connected with internet is must as it communicates with server

 Networks issues have impact on end results

7.3 Applications

 Precision Farming

Dept of ECE,GMIT,Davangere 2022-23

45
 Pick and Place Movable Machine

Chapter 8
CONCLUSION & FUTURE SCOPE

The suggested project is an effort to apply the machine learning idea to crop recommendation via the
implementation of KNN and Nave Bayes algorithms, as well as other techniques. It is necessary to
gather and utilize real-time soil test reports from the district agricultural department as training data.
Testing data is data that is directly obtained from the soil grids via the REST API (Representational
State Transfer). After normalizing the raw data, the Knn algorithm for crop selection and the Naive
Bayes method for yield prediction were applied. The whole process is represented visually via the
use of Google's Bar Chart API. It is hoped that it would be helpful in precision agriculture, which is
becoming more popular in India.

 Future scope

The project may be further enhanced by increasing the number of observations, such as soil test data,
and it could also be implemented by including more machine learning algorithms.

Dept of ECE,GMIT,Davangere 2022-23

46
 Pick and Place Movable Machine

REFERENCES

1] S. Pudumalar, E. Ramanujam, R. HarineRajashree, C. Kavya, T. Kiruthika, J. Nisha depicts

the Yield Recommendation System for Precision Agriculture.
This author introduces the fundamentals of precision farming before moving on to developing a
model. On the rancher and product levels, it shows the accuracy of horticulture needs for small, open
ranches. In India, the most common issue that farmers encounter is that they do not pick the most
suitable crop for their particular soil type. As a consequence of this, they will see a substantial
decrease in their output. We want to reach wish to do this by using the most cutting-edge
technologies, such as SMS and email, as the primary destinations for advice warning
administrations, in order to reach even the tiniest rancher at the level of his or her smallest harvest
plot. This model was created expressly for the state of Kerala, where typical holding sizes are much
less than those of the rest of India.
2] Saiyyad Mohmmad Ali, Muzffar Ali depicts Study of order calculations for detailing yield
forecast exactness in accuracy
Agriculture planning is essential in every nation, and it is especially significant in developing
countries. The agriculture sector generates a variety of products, including food, raw materials for
industry, economic stimulus, and job creation. There is a wealth of information available on the
variables that influence the input and output of the agricultural industry. As technology progresses,
new data mining methods are being developed and implemented. These data mining methods may be
used to examine the multidimensional, time specific data of the agricultural sector in order to get
useful information from it that can be used to help the economy grow. Throughout this paper, we
will examine various approaches taken by different researchers to deal with the analysis of the
agriculture sector in order to provide an effective way to increase production while also making the
most of available resources, which in turn will increase the economy in this sector.
3] Supriya D M Data Mining Approach for Soil Behavior Analysis and Crop Yield Prediction
Yield prediction contributes to the proper selection of crops for sowing which is very popular
among farmers these days [4]. An interesting challenge is predicting and yielding of the crops. In
earlier days yield prediction was performed by considering the farmers experience on a particular
field and crop. Yield prediction uses the data mining techniques 4 in 4 order 2 to 9 predict the
category of the analyzed soil datasets. Yielding of crops will be indicated by prediction Naïve
Bayes and K-Nearest Neighbor methods are used for predicting the crop yield.

Dept of ECE,GMIT,Davangere 2022-23

47
 Pick and Place Movable Machine
4] E. Manjula, S. Djodilta Choum depicts An Agricultural Yield Prediction Model.
In prior days yield expectation was performed by considering agriculturists encounter on specific
field and product [5]. One of the significant issues that remaining parts to be fathomed in light of
accessible information is yield forecast. One of the better decisions for this is information mining
systems. For assessing what's to come year's harvest generation distinctive Data Mining systems are
utilized and assessed in farming. A framework to anticipate trim yield from past information
proposed and executed in this exploration. It can be accomplished by applying affiliation administer
mining on farming information. Formation of a forecast demonstrate which might be utilized to
future expectation of harvest yield is focused in this exploration. In this paper it shows a short
examination of product yield expectation utilizing information mining strategy in view of affiliation
rules for the chose locale i.e., locale of Tamil Nadu in India.
5] Anshal4Savla, Himtanaya*Bhadada, ParulDhawan, Vatsa Joshi depicts the use of machine
learning techniques for yield prediction on delineated zones in precision agriculture.
Normalization is the process through which information in a database is sorted and organized. There
are a lot of stages involved. A large quantity of data is gathered in agriculture, and a variety of data
mining methods are used to make effective use of this information. It is possible to diffract a table
into a lot of smaller tables that are less repetitious. Outside keys are defined in old tables that refer to
the important keys in new tables, and no data is lost in the course of the process. The information
standardization process has been completed in order to disconnect it. In order for the progressions
made in the characteristics of one table to be communicated to the whole database as efficiently as
possible. From the perspective of data mining, this boils down to a variation of spatial clustering that
is constrained by the requirement that the resultant clusters be geographically largely continuous in
nature. The ultimate aim is to guarantee that the information generation of tables and the
characterization of relationships between them is completed while keeping this in mind. Data
normalization is a helpful idea in order to arrange a large data collection of information. The data is
divided into intelligent gathers, with each gathering correctly depicting the little bit of information
that was originally collected. It is possible to modify the information in a large database efficiently.
By adjusting the date range to be more consistent It has been accomplished to make information
more accessible and to manage it more quickly.

6] Suma, N., Samson, S. R., Saranya, S., Shanmugapriya, G., & Subhashri, R. (2017)
IOT based smart agriculture monitoring system. International Journal on recent and Innovation
Trends in computing and communication, 5(2), 177-181.

Dept of ECE,GMIT,Davangere 2022-23

48
 Pick and Place Movable Machine
7] Prathibha, S. R., Hongal, A., & Jyothi, M. P. (2017, March)
IOT Based monitoring system in smart agriculture. In 2017 International Conference on Recent
Advances in Electronics and Communication Technology (ICRAECT) (pp. 81-84). IEEE. G.
8] Ramkumar, P. Parkavi, K. Ramya and M. S. Priya
"A Survey on Sar Images Using Image Processing Techniques,” 2020 6th International Conference
on Advanced Computing and Communication Systems (ICACCS), 2020, pp. 1097-1100, doi:
10.1109/ICACCS48705.2020.9074261.
9] Chandoul Marwa, Soufiene Ben Othman and Hedi Sakli
"IoT Based Low-cost Weather Station and Monitoring System for Smart Agriculture", International
Conference on Sciences and Techniques of Automatic Control and Computer Engineering, 2020.
Mohamed Rawidean Mohd Kassim, “IoT Applications in Smart Agriculture: Issues and Challenges",
IEEE Conference on Open Systems, 2020 G.
10] Ramkumar, A. G, S. K. M, M. Ayyadurai and S. C
"An Effectual Underwater Image Enhancement using Deep Learning Algorithm," 2021 5th
International Conference on Intelligent Computing and Control Systems (ICICCS), 2021, pp. 1507-
1511, doi: 10.1109/ICICCS51141.2021.9432116

Dept of ECE,GMIT,Davangere 2022-23

49
•