SDM COLLEGE OF ENGINEERING AND TECHNOLOGY,

Dharwad-580002

(An autonomous Institution af liated to

Visvesvaraya Technological University, Belgavi – 590018)

Department of Electrical and Electronics Engineering

A Synopsis On The Major-Project Entitled

“Improved Plant Leaf Disease Detection using Machine Learning ”

SUBMITTED BY
Name of the student USN
Manjunath H S 2SD21EE024

Pavan Raykar 2SD21EE031

Prerana S Panchal 2SD21EE034

Varadayini Patil 2SD21EE059

th
Students of 7 Semester.

Under the guidance of

Prof. V R SHEELAVANT
Department of EEE,
SDM College of Engineering & Technology,
Dharwad

Academic Year: 2024-2025

fi
 TABLE OF CONTENTS
Contents Page Number

Introduction 03

Rationale 03

Problem statement 03

Objectives 03

Literature survey 04

Methodology 04

Components 07

Expected outcomes & References 07

2

 1. Introduction

Agriculture is vital for providing food and combating climate change. However, plant diseases pose
significant threats to agricultural productivity. While digital image processing tools are effective for
identifying plant diseases, human observations often fall short. Neural network techniques like Back
propagation and Principal Component Analysis (PCA) have been employed for fungal disease
identification. This project aims to explore various plant leaf diseases and evaluate different image
processing techniques for accurate disease detection. The system will be capable of identifying
seven diseases, including Down Mildew, Early Blight, Mosaic Virus, Leaf Miner, and White Fly.
We will specifically investigate the use of the K-Nearest Neighbours (KNN) classifier, a promising
alternative to SVM for multi-class classification tasks.

Rationale

The proposed project aims to leverage the power of digital image processing and k-nearest
neighbours (K-NN) to revolutionise plant disease detection. By employing K-NN, the system will
classify plant leaf images based on their similarity to previously labeled examples. This approach
offers a relatively simple yet effective method for disease identi cation, as it leverages the inherent
patterns and relationships within the image data. The system will be trained on a dataset of labeled
images, allowing it to accurately classify new/unseen images.

1.1 Problem Statement

Every other field has got some benefit from new technologies as compared to the agricultural field.
According to past studies, 42% of agricultural production is in loss and that too only because of the
increasing rate of loss due to plant leaf diseases. To overcome this major issue, this plant leaf
disease detection technique can be applied to detect a disease from the input images. This process
involved steps like image preprocessing, image segmentation, feature extraction.

1.2 Objectives

• To find the interactions between the disease causing agent and host plant in relation to overall
environment.

• To identify various diseases in plant’s leaf.

• To implement a method for preventing the diseases and providing solutions for reducing the
losses/damages caused by diseases.

3

fi
 2. Literature Review

1. "A Review of Deep Learning Techniques for Plant Disease Detection" (2019) by S.
Mohanty et al.

• This paper provides a comprehensive overview of deep learning techniques for plant
disease detection. It discusses various architectures like CNNs, RNNs, and hybrid
models, and highlights their strengths and limitations.
2. "Plant Leaf Disease Detection Using Deep Learning: A Review" (2020) by S. Karthik et
al.

• This review focuses on deep learning-based methods for plant leaf disease detection. It
covers different datasets, pre-processing techniques, and evaluation metrics used in the
eld.
3. "A Review of Machine Learning Techniques for Plant Disease Diagnosis" (2021) by A.
Kumar et al.

• This paper explores both traditional machine learning and deep learning approaches for
plant disease detection. It compares the performance of different algorithms and discusses
challenges in implementing these techniques.
4. "Plant Disease Detection Using Deep Learning: A Survey" (2022) by M. Sharma et al.

• This survey provides a detailed analysis of deep learning-based methods for plant disease
detection. It covers various architectures, feature extraction techniques, and transfer
learning approaches.
5. "A Systematic Review of Plant Disease Detection Using Deep Learning" (2023) by P.
Singh et al.

• This review focuses on the recent advancements in deep learning for plant disease
detection. It covers state-of-the-art techniques, challenges, and future research directions.

3. Methodology

4

fi
 1. Image Acquisition:

• Use the Raspberry Pi to capture real-time images of plant leaves.

• Ensure adequate lighting and focus for clear image quality.
2. Pre-processing:

• Convert RGB images to grayscale.

• Resize images to a standard size for consistent processing.
• Normalise pixel values to a common range.
3. Feature Extraction:

• Extract texture features using Gray-Level Co-occurrence Matrices (GLCM).

• Consider additional features like colour histograms or shape descriptors if necessary.
4. Classi cation:

• Employ a machine learning algorithm (e.g., Support Vector Machines (SVM), K-

Nearest Neighbours (KNN)) to classify images based on extracted features.
• Train the classi er on a labeled dataset of plant leaf images.

5. Real-time Analysis:

• Integrate the classi cation model into a real-time application on the Raspberry Pi.
• Process captured images to detect potential diseases.

Research Type

The project aims to develop a real-time plant disease detection system using image processing and
machine learning techniques.

Tools of Data Collection / Analysis

Software: The Matlab program.

Hardware :
• Raspberry Pi: A low-cost, single-board computer for image acquisition and
processing.
• Camera: A suitable camera (e.g., USB webcam, Raspberry Pi camera module)
for capturing plant leaf images.

5

fi
fi
fi
 4. Software/Hardware Required
Hardware:

• Raspberry Pi: A single-board computer for image processing and analysis.

• Camera: A USB webcam or Raspberry Pi camera module for capturing plant leaf images.
• SD Card: A microSD card to store the operating system and data.
• Power Supply: A suitable power adapter for the Raspberry Pi.
Software:

• Programming Language: Basic Matlab Software for image processing.

5. Expected Outcomes

This project include achieving high accuracy rates in disease classification, ensuring real-time
processing capabilities, and contributing to improved agricultural practices. By leveraging image
processing techniques and machine learning algorithms, the system aims to accurately identify
various plant diseases, thereby enabling timely intervention and reducing crop losses. Additionally,
the project seeks to develop a system that is user-friendly and easily adaptable to different
agricultural settings, promoting its widespread adoption and contributing to enhanced agricultural
sustainability.

References

• Sladojevic, S., Arsenovic, M., Culibrk, D., & Anderla, A. (2016). Deep learning for leaf disease
detection: A review. Computers and Electronics in Agriculture, 131, 14-22.

• Mohanty, S. P., Hughes, D. P., & Salathé, M. (2017). Using deep learning for image-based plant
disease detection. Frontiers in Plant Science, 8, 154.

• Zhang, D., Wang, G., Wang, Y., & Li, H. (2018). Plant disease detection using deep convolutional
neural networks. Computers and Electronics in Agriculture, 153, 142-150.

• Kamble, S., & Patil, P. (2019). Plant leaf disease detection using image processing and machine
learning. International Journal of Engineering Research & Technology, 8(5), 1585-1590.

• Khan, M. A., & Lee, J. (2020). Plant disease detection using deep learning techniques. Computers
and Electronics in Agriculture, 174, 105764.

6

 Name of the student USN Signature
Manjunath H S 2SD21EE024

Pavan Raykar 2SD21EE031

Prerana S Panchal 2SD21EE034

Varadayini Patil 2SD21EE059

GUIDE NAME SIGNATURE

Prof. V R Sheelavant

7