International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056

Volume: 09 Issue: 04 April 2022 www.irjet.net p-ISSN: 2395-0072

Plant Leaf Disease Detection Using Machine Learning

Jayshree Hajgude1, Jayesh Kriplani2, Dhiraj Chhabria3, Anish Verliani4

1 Professor, Dept. of Information Technology, VESIT, MUMBAI

234 Student, Dept. of Information Technology, VESIT, MUMBAI
---------------------------------------------------------------------***------------------------------------------------------------------
Abstract— The Project presents plant leaf disease image, and if the leaf has a disease, the name of the disease,
detection, effect of disease on plant yield and what as well as the prescribed pesticides, will be presented on
pesticides should be used for its cure. In agriculture, the user's screen after pressing the predict button. If there
research of automatic plant disease is essential one in is no disease in the plant leaf after uploading the image, a
monitoring large fields of plants, and thus automatically message reading "There is no disease on the Plant" will be
detects symptoms of disease as soon as they appear on displayed. The system also displays the proportion of
plant leaves. impacted areas and recommends pesticides based on that
percentage.
Our research aims to solve the problem of detecting and
preventing diseases of agricultural crops.. To determine Need of plant leaf disease detection:
the optimal architecture for deep learning, we considered
several models. As a source of the training data, we use the
PlantVillage open database For this approach automatic We have all seen how the farmers as well as all the people
classifier Convolutional Neural Networks (CNN) model will get Infected due to distortion of plants due to various
be used for classification based on learning with some diseases. This problem not only affects the farmers income
training samples. The developed model is deployed as a but it also affects the consumers due to shortage of supply.
web Application which can detect 15 types of diseases The estimated loss due to plant leaf disease is 60 billion
among plants viz. Tomato, Potato and Pepper. dollars which is a huge amount of money.

Validation Accuracy Results show that the system can One of the most important needs of this project is the ease
show the accuracy 93.50 % and Training Accuracy is of detection of the disease by the farmers in a very simple
approximately 97.87% manner as farmers cannot predict the disease just by
looking at the leaf. This system will also help farmers
Key Words : plant leaf disease, Plant Village Dataset , identify the disease and also recommend certain pesticides
pesticides, CNN, Accuracy. which they can use to cure the disease and stop further
spread of the disease as if one plant leaf is Infected there is
a chance that disease might spread on many leaves and
I. INTRODUCTION destroy the Total yield.

Plants have become a major source of energy and are a PROBLEM STATEMENT
critical component in the solution to the global warming
crisis. Several plant diseases have the potential to bring
disastrous economic, social, and ecological consequences. To make an efficient use of Machine Learning Algorithms
To meet projected demand, worldwide crop production which reduces time and cost for Farmer to detect the plant
must increase by at least 50% by 2050. disease, its effect on crop yield and suggest the pesticides
for Plant disease.
Currently, the majority of production takes place in Africa
and Asia, where 83 percent of growers are family-run II. SCOPE
businesses with little to no horticultural experience. As a
result, yield losses of more than 50% are common as a This project will be very helpful to farmers in rural areas
result of pests and diseases. The old approach of human and also will help them in saving their yields from diseases
analysis by visual inspection is no longer viable for as farmers lose a huge amount of their cultivated crops
categorizing agricultural diseases. because of diseases and this system will help them to avoid
the similar situation. Also we have tried to implement this
The main goal of this project is to identify the most project in a regional language so as to make things more
common diseases seen in Plant Leaf. Tomato, potato, and understandable for farmers.
pepper are the three types of plants employed in this
research. This system can detect 15 different forms of This project can also provide them cure by suggesting
illnesses in these three plants. The user can upload a leaf pesticide names and their quantity according to the

© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 3113
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 04 April 2022 www.irjet.net p-ISSN: 2395-0072

prevailing disease and this project will also display and model and the Plant Village dataset is used for model
give an idea about the effect of crop disease on crop yield. training and validation of the entire system.

III. RELATED WORK Now there are two possibilities after uploading the image:
either the plant leaf will be infected or the plant leaf won't
be Infected. If the leaf is found to be infected the image of
There are different algorithms and methodologies for the leaf will be displayed on screen along with the name of
identifying the disease on plant leaves. There are many the disease in English and Hindi language. Also the
different organizations and many researchers who have proposed system is capable of calculating the percentage
studied and have done work on this topic using different area Infected and displaying it on the screen and also
algorithms. Some of them are summarized below. suggesting the pesticides on the basis of the area Infected.
The name of the pesticide, image of the pesticide is
In paper [1], In 2018 in the International Journal of Science displayed along with the quantity and all of the
and Engineering Applications Mr.Ko Ko Zaw, Dr. Zin Ma Ma information is displayed in two languages English and
Myo,Mr. Daw Thae Hsu Thoung proposed Image Hindi for better understanding of farmers as many of the
processing is the only way to build the simple, robust and farmers won't be able to understand name of disease in
accurate disease detection system While working with English but the farmers would be familiar with regional
image processing. Detailed studies should be performed on hindi names of the disease that is present on plant leaf.
the types of diseases, their symptoms, and the patterns of
disease. The system will be designed based on the patterns Alternatively, if the leaf is not infected, the image along
of disease. The mainly occurring diseases on leaves are with the message "There is no disease in the leaf of the
Bacterial disease, Fungal disease, Viral disease and plant" is displayed.
diseases due to insects. These diseases are detailed in the
paper. B. Proposed Methodology

In paper [4], In 2018 in International Conference on Design Plant leaf disease detection system flowchart:
Innovations for 3Cs Compute Communicate Control Shima
Ramesh, Mr. Ramachandra Hebbar, Niveditha M, Pooja R,
Prasad Bhat N, Shashank N,Mr. P V Vinod proposed
According to past studies, 42% of agricultural production
is in Loss, and that too solely as a result of plant leaf
diseases. With this technique, plant leaf diseases can be
detected from input images to overcome this major issue.
This process involved steps like image preprocessing,
image segmentation, feature extraction. On the basis of
these three steps, K Nearest Neighbor (KNN) classification
is applied. As a result of the proposed implementation, A
98.56% accuracy rate in predicting plant leaf diseases.

In paper [5], In 2021 in BMC(Biomedcentral) Jun Liu and

Xuewei Wang proposed plant disease and pest detection
using SVM algorithm. SVM is used for recognizing image
samples. The accuracy of implementation is about 92%
with a sufficient and stable dataset used for the project.

IV. PROPOSED SYSTEM

A. Introduction

The first step is the farmer will upload the image of the leaf
on the website. After uploading the image, farmers can
click on the predict button and wait for the result. For this
CNN algorithm , the keras module is used to train the

© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 3114
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 04 April 2022 www.irjet.net p-ISSN: 2395-0072

1. Home Page:
This is the first page of the website.

4. If the plant leaf has any disease, this page will

show the name of the disease with the image of
the leaf and upto how much extent the leaf is
Infected. For eg. Here the plant leaf has ‘Potato
Early Blight’ disease with 23.415% region Infected
which is shown here.

2. In this page, the user will have to upload the image

of the leaf and the user have to click on the
‘Predict’ button to predict the leaf disease.

5. After predicting the disease, this page will

recommend the products through which the plant
leaf disease can be cured.

3. After uploading the image of a leaf in the previous

page and after clicking on predict if the plant does
not have any disease the page will show that there
is no disease.

© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 3115
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 04 April 2022 www.irjet.net p-ISSN: 2395-0072

V. CONCLUSION

To prevent losses, small holder farmers are dependent on a

timely and accurate crop disease diagnosis. We fine-tuned
a pre-trained Convolutional Neural Network, and then
deployed it online. The final result was a plant disease
detection app. This service is free, easy to use and requires
just a smartphone and internet connection. Overall, this
study is conclusive in demonstrating how CNNs may be
applied to empower small-holder farmers in their fight
against plant disease. In the future, work should be focused
on diversifying training datasets and also in testing similar
web applications in real life situations. Without such
developments, the struggle against plant disease will
continue.
6. Below image shows the Infected parts of a leaf.
So Implementation of the project work was done
successfully and we got to learn many more things from
this project.

VI. ACKNOWLEDGEMENT

This project report would not be successful and final

without the guidance and assistance of many people, and
we are very fortunate to have received this during the
completion of our project. Nothing we have accomplished
is possible without their support and assistance, and all we
have accomplished is only because of such support and
assistance.

It is with great respect and gratitude that we extend our

thanks to Dr. Shalu Chopra, Head of the Department of
Information Technology, for granting us this permission to
perform this work. We are incredibly grateful to her for the
nice support and guidance she provided to us throughout
this project, in spite of her busy schedule managing the
department.
7. Displaying the total and infected area of a leaf. And
also Percentage of infected regions. Taking great interest in our project work and guiding us all
along, till completion of our project work, we owe our
deepest gratitude to our project guide, Mrs. Jayshree
Hajgude, who helped us to develop a good report by
providing all of the necessary information. Also, we would
like to thank all the authors named in the literature survey
as well as the references section for sharing their
discoveries with us. Getting to our goal would be virtually
impossible without their contributions.

VII. REFERENCES

1. The detection and classification of leaf diseases

using Multiclass Support Vector Machines, 11th
National Conference on Science and Engineering,
Yangon, Yangon, Myanmar, Ko Ko Zaw, Zin Ma Ma
Myo, and Wah Wah Hlaing.

© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 3116
 International Research Journal of Engineering and Technology (IRJET) e-ISSN: 2395-0056
Volume: 09 Issue: 04 April 2022 www.irjet.net p-ISSN: 2395-0072

2. A Survey of Disease Identification and

Classification Through Leaf Images:A
Computational MethodSukhvir Kaur, Shreelekha
Pandey, Shivani Goel.Received: 26 June 2017 /
Accepted: 11 January 2018 CIMNE, Barcelona,
Spain 2018

3. Shima Ramesh Assistant Professor: department of

electronics and communication, MVJ College of
Engineering. Bangalore, India, 2018 International
Conference on Design Innovations for 3Cs
Compute Communicate Control Plant Disease
Detection Using Machine Learning.
Niveditha M, Pooja R, Prasad Bhat N, Shashank N,
Research Scholar: department of electronics and
communication, MVJ college of Engineering,
Bangalore, India.

4. The detection of plant diseases and pests using

deep learning: a review by Jun Liu and Xuewei
Wang Plant Methods (2021) 17:22

5. 2017 IEEE International Conference on

Technological Innovations in ICT For Agriculture
and Rural Development Identification of Plant leaf
disease using image processing. Pooja V, Rahul
Das, Kanchan V, Department of Computer
Science,Amrita School of Arts and Sciences, Amrita
VishwaVidyapeetham, Amrita University, Mysuru
Campus, Karnataka, India.

6. International Journal of Science and Engineering

Applications Volume 7–Issue 08,143-147, 2018,
ISSN:-2319–756 . Support Vector Machine Based
Classification of Leaf Diseases - Ko Ko Zaw,
Department of Electronic Engineering
Technological University, Thanlyin, Myanmar
Dr. Zin Ma Ma Myo, Department of Electronic
Engineering Technological University, Thanlyin,
Myanmar

© 2022, IRJET | Impact Factor value: 7.529 | ISO 9001:2008 Certified Journal | Page 3117