18CSP401L - PROJECT WORK FINAL REVIEW

PRESENTATION

Date & Session :12/04/2023 & AN

Batch Number : 28

1
Name of the Candidates : Megha M
Roshni M

Register Numbers : 19BCS4073

19BCS4103

Name of the Supervisor : Karthika I , AP/CSE

with Designation

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 2

 Leaves of a plant can be used to determine the health status of that plant.
 A deep-learning-based approach is presented for the plant disease detection and
classification from leaf images captured in various resolutions.
 The objective of this work is to develop a system that capable to detect and identify
the type of disease based on a Classification algorithm.

18CSP401L - PROJECT WORK FIANL REVIEW PRESENTATION 3

 Automate this process to show the prediction result in Web Application or Desktop
Application.
 To optimize the work it should be implemented in Deep Learning Algorithm .

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 4

 Early plant disease detection plays a significant role in efficient crop yield.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 5

 In India, people are most likely interested in agriculture. It plays a very important
role in the Indian economy.
 Recent days there is many types of insects that affects the growth of the crops and
damages the fruits, leaf and flowers.
 Various insects are grasshoppers, weevils, beetles and small insects like fly larva
and thrips can cause serious damage to crops.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 6

S.No Authors Title of the paper Methodology Merits De Merits
& Years
S. V. A comprehensive review on Machine learning and Deep Robust and easily It takes more time to
1
Kogilavani, S. detection of plant disease learning deployable detect disease
Malliga in using machine learning and strategy effectively as it takes
2022 deep learning approaches long process

Sunil S. Plant leaf disease detection Support Vector Machine Transfer Learning It takes more time to
2
Harakannanav using computer vision and (SVM), (CV) and K- accuracy is better training the dataset
ar and et al in machine learning algorithms Nearest Neighbor (K-NN). than traditional depends upon the
2022 model. quality of images.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 7

S.No
Authors Title of the paper Methodology Merits De Merits
& Years
P. Deepika , S. Pest and Disease Monitoring Minimum Distance Disease is identified in To improve
3
Kaliraj,et.al in of Crops Classifier MDC,GLCM, initial stage and uses recognition rate in
2021 Minimum Distance Criterion less computational classification other
and SVM effort models can be used.

Mustafa AI-based Detection of Pest Convolutional neural To improve accuracy, Produces a very low
4
Ahmed, Tushar Infected Crop and Leaf networks (CNN) mixed variety of accuracy for the
Mahajan,et.al in images are loaded in images taken in real
2021 training phase. environment.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 8

 Limited availability of large-scale datasets for training and testing machine
learning models
 Need for more efficient and automated methods that can accurately detect and
classify diseases in real-time
 Limited research on other machine learning algorithms beyond deep learning for
plant leaf disease detection

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 9

Dataset Creation CNN Pre-
Training Model
Train , Test , Val Design Processing

Load Model Pre-Process

Classification
Data Input Image

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 10

❑ Hardware Requirements
• Operating System: Windows/Linux
• Simulation Tool: Pycharm /Anaconda jupyter
❑ Software Requirements
• Processor: AMD RADEON/intel i5
• Hardware: Minimum 80 GB
• RAM: Minimum 2GB

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 11

❖ Module 1: Training Module
❖ Module 2: Pre-Processing
❖ Module 3: Testing Module
❖ Module 4: Classification

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 12

TRAINING MODULE
 Feature Extraction extracts all the images of leaf i.e. input.

 Feature selection performance disease classification from the disease database.

 Disease database contain all the various diseases of the leaf.

 Thousands of different leaf dataset will be trained.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 13

PRE-PROCESSING MODULE

 Pre-processed images are reduced image size and image crop to a given input.
 The study uses colored and resized images to 96x96 resolution for processing.
 A Convolutional Neural Network (CNN) is a Deep Learning algorithm which can
take in an input image, assign importance to various objects in the image and be
able to differentiate one from another.
 The pre-processing required in a CNN is much lower as compared to other
classification algorithms.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 14

TESTING MODULE
 The purpose of testing is to evaluate the performance of the model on a separate
dataset that the model has not seen before.
 By using deep learning it compares the diseases from the feature selection
classification to find the type of disease that leaf has.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 15

CLASSIFICATION MODULE
 Classification uses a fully connected layers and for feature extraction it uses
convolutional and pooling layers.
 The classification process classifies the plant leaf if it is infected with the disease
or not, identifies the type of plant disease and recognize the plant variety.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 16

MAIN STEPS TO BUILD A CNN
1.Convolution Layer
It is the first layer to extract features from the input image and it learns the
relationship between features using kernel or filters with input images.
2. ReLU Layer
ReLU stands for the Rectified Linear Unit for a non-linear operation. The
output is f(x) = max(0,x). we use this because to introduce the non-linearity to CNN.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 17

3. Pooling Layer
It is used to reduce the number of parameters by down sampling and retain
only the valuable information to process further.

➢ Max Pooling

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 18

4.Fully Connected Layer
It helps to pass our flatten vector into input Layer. We combined these
features to create a model. Finally, we have an activation function such as softmax to
classify the outputs.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 19

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 20
18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 21
▪ In rural areas farmers spent hours on examining every crop for their health. This
method is not a suitable and accurate method.
▪ Deep learning techniques are used for accurate prediction of diseases .
▪ It gives better accuracy than any other traditional methods.
▪ Detection accuracy can be increased by training the model with more dataset.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 22

▪ A.Umamageswari, S. Deepa, K. Raja, “ An enhanced approach for leaf disease
identification and classification using deep learning techniques – 2021”.
▪ Majji V Applalanaidu, G. Kumaravelan , “ Machine Learning Approaches in Plant
Leaf and Classification – 2021”.
▪ Sherif E. Abdelhamid, Roxane Elias Mallouhy , Jaafar Alghazo , “ Detection of
Rice Plant Diseases Using an Improved CNN Model – 2022”.
▪ Shoyeb Ahammad Rafi, Toukir Hasan Chowdhury, Dewan Md. Anisur Rahaman,
Md. Washimul Bari, Md. Jahidul Islam, “ Crop Condition Analyzer Robot with
Image Processing Feedback to Assist Farmer-2021”.
18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 23
 S. Nandhini, Dr K. Ashokkumar , “ Analysis on Prediction of Plant Leaf diseases
using Deep Learning – 2021”.
 R Vijaya Saraswathi , Vaishali Bitla, P Radhika,T Navneeth Kumar, “ Leaf Disease
Detection and Remedy Suggestion Using Convolutional Neural Networks – 2021”.
 Mariappan Sakkan, Sanat Sarangi, Swagatam Bose Choudhury, Prachin Jain,
Hemavathy B., Rajesh Urkud, “ Disease Assessment in Rice with Pervasive
Stationary and Mobile Sensing – 2021”.

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 24

I.Karthika , M.Megha, M.Roshni, “ Deep Learning Approach to Automated
Tomato Plant Leaf Disease Diagnosis”. Second International Conference on
Electronics and Renewable Systems(ICEARS-2023),St. Mother Theresa
Engineering College, Tuticorin (Accepted for Publication in Lecture Notes on
Electronics and Renewable Systems, IEEE ).

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 25

18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 26
18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 27
18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 28
Thank you
18CSP401L - PROJECT WORK FINAL REVIEW PRESENTATION 29