Scribd
Upload
280 views29 pages

Leaf Disease Detection Using Decision Tree

This document presents a methodology for detecting leaf diseases using decision trees. It aims to automatically detect disease symptoms in plant leaves by analyzing digital images. The methodology includes pre-processing images through binarization and filtering, segmenting images using connected component labeling, extracting features using gray-level co-occurrence matrices, and classifying diseases using decision trees and convolutional neural networks. The results show the model can accurately detect different diseases like bacterial spot, late blight, and fungal diseases in various plant leaves like banana, beans, and lemon. Future work could focus on improving algorithms to more rapidly and effectively identify diseased leaves.

Uploaded by

Power Bang
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
280 views29 pages

Leaf Disease Detection Using Decision Tree

This document presents a methodology for detecting leaf diseases using decision trees. It aims to automatically detect disease symptoms in plant leaves by analyzing digital images. The methodology includes pre-processing images through binarization and filtering, segmenting images using connected component labeling, extracting features using gray-level co-occurrence matrices, and classifying diseases using decision trees and convolutional neural networks. The results show the model can accurately detect different diseases like bacterial spot, late blight, and fungal diseases in various plant leaves like banana, beans, and lemon. Future work could focus on improving algorithms to more rapidly and effectively identify diseased leaves.

Uploaded by

Power Bang
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PPTX, PDF, TXT or read online on Scribd
You are on page 1/ 29

LEAF DISEASE DETECTION

USING DECISION TREE


FATIMA MASOOD
160617742120
INTRODUCTION
►India is an agriculture dependent country.

►70% of Indian economy depends on agriculture , but leaf infection phenomena


causes the loss of major crops results in economic loss.

►In plants, diseases usually occur on leaves, fruit, buds, and branches.This
situation causes plants to be damaged.

►For this reason, it is very important to determine the disease in advance and to
take the necessary precautions before it spreads to other plants.
OBJECTIVE

►The main objective of this project is to detect the unhealthy region of plant
leaf's using machine learning.

►To detect leaf disease portion from image.

►To extract features of detected portion of leaf.

►To recognize disease from detected portion of leaf through Machine learning
algorithms i.e. CNN and Decision tree algorithms.
PROBLEM DEFINITION
During cultivation process, Leaves expose to many problems & diseases such as :
Late blight
Anthracnose
Yellow curl virus
Bacterial spot
The naked eye observation of experts is the main approach adopted in practice for
detection and identification of plant diseases . But this approach requires continuous
monitoring of experts which might be expensive and difficult especially in large
farms.
PROBLEM DEFINITION (CONTI.)

So, it is necessary to help farmers in automatically detect symptoms of disease


as soon as they appear by analysing the digital images which may helping us
for:
Minimizing major production & economic losses.
Ensuring both quality and quantity of agricultural products.
Minimizing agrochemicals use.
MOTIVATION

• India is well known for its agriculture production.


• Farmers have variety of options to cultivate crops in the field. Still, the
cultivating these crops for best harvest and top quality of production is done in
a technical way . So, the yield can be increased and quality can be improved
by the use of technology.
• Generally, whenever there is disease to plant, we can say that leaves are the
main indicator of the disease caused to plant.
LITERATURE SURVEY
Author Year of Description Pros Cons
publication

Muammer and 2019 Plant disease and pest detection This is used for leaf detection by Recognition rate can be
Davut using deep learning. using deep neural network for more effective.
deep feature extraction.

Piyush chaudhry 2018 Color transform based approach Median filter is used for mage Disease spot area can be
for disease spot detection on plant smoothing & threshold can be computed for assessment
leaf. calculated by applying OTSU of loss in agriculture crop.
method.

Sabah bashir 2017 Remote area plant disease Texture segmentation by Bayesclassifier, k mean
detection using image occurrence & k means clustering clustering & principal
processing . technique. component classifier can
be used to classify various
plants.

Anand H. 2016 Applying image processing Gabor filter for feature Recognition rate can be
kulkarni technique to detect plant extraction & ANN classifier for increased.
diseases . classifications
METHODOLOGY
Input image

Pre Processing

Segmentation

Feature Extraction

Split

Train Test

Model Evaluation

Result
DATASET

• The dataset used here is collected from Kaggle “leaf disease Dataset”.
• It contains over 2000 image samples of jpg files(256x256).
• It includes different diseases like: Bacterial Spot, Late Blight, Anthracnose,
yellow curl virus.
S.NO Type of Disease Count
1 Healthy 345
2 Late Blight 700
3 Bacterial Spot 400
4 Yellow Curl Virus 355
5 Anthracnose 200
PRE-PROCESSING
Binarization
• Input images are of from different writers and may be in different colours, it needs
to be binarized so that it may be used for experimental use.
• The Document image is a grayscale image. One may convert this image into
binary/ two tone image which is having only two levels of intensity.
• The threshold value used in binarization is 127.
• The Otsu’s thresholding method is used for binarization of the image.
Morphological Filtering

 Opening operation is a filtering method which removes all the objects that
contains fewer than 30 pixels of the input binary image and hence
imperfections are removed. The document image may be contaminated with
noise so filter is applied to remove any spurious information.

 The Image after cleaning is subjected to segmentation of the image. The


opening operation is used as a morphological operation in this work.
SEGMENTATION
Connected Component Labelling
• Connected components labelling scans an image and groups its pixels into
components based on pixel connectivity, i.e. all pixels in a connected
component share similar pixel intensity values and are in some way
connected with each other.
• Connected component labelling is done before segmentation. Labelling is just
giving a pixel a particular value. The connected component labelling
algorithm consists of two passes.
FEATURE EXTRACTION
Gray-Level CO-occurence Matrix
Geometric method of examining texture that considers spatial relationship of pixels
is the GLCM. GLCM function characterize the texture image by the pairs of pixel
with specific values and specific spatial relationship occur in an image.
In Grayco-matrix, to calculate pixel size intensity value i occur in a specific spatial
relationship to a pixel with the value j. each element (i,j) in GLCM specifies the
number of times pixel value i occurred adjacent to a pixel with value j.
Adjacency can be specified to occur 4 direction 0°,45°,90°,135° in 2-d square pixel
image.
 GLCM stores instance occurrences between adjacent pixels. According to the
definition of GLCM, the co-occurring pairs obtained by choosing θ . Hence, one has
four choices to select the value of θ.
MACHINE LEARNING CLASSIFIER
Decision Tree
• Decision Trees are a non-parametric supervised learning method used for
both classification and regression tasks.
• The goal is to create a model that predicts the value of a target variable
by learning simple decision rules inferred from the data features.
• A decision tree is a graphical representation of specific decision situations that are
used when complex branching occurs in a structured decision process.

CNN(Convolutional Neural Network)


• It is a specific type of artificial neural network that uses perceptrons, a machine
learning unit algorithm, for supervised learning, to analyze data.
• CNNs apply to image processing, natural language processing and other kinds of
cognitive tasks.
RESULTS

• Input and output image of banana leaf and output diseases is early
scorch disease.
RESULTS

• Input and output image of beans leaf and output diseases is fungal disease.
RESULTS

• Input and output image of beans leaf and output diseases is bacterial leaf spot.
RESULTS

• Input and output image of lemon leaf and output diseases is sun burn disease.
EXPERIMENTAL RESULTS

Detection Accuracy
Detection Accuracy for detecting the exact
disease.
CONCLUSION

This project gives the information on different diseases classification


techniques used for plant leaf disease detection and an algorithm for image
segmentation technique that can be used for automatic detection as well as
classification of plant leaf diseases later.
Another advantage of using this method is that the plant diseases can be
identified at early stage or the initial stage.
Pre-processing of the images is done in two steps i.e. binarization and
morphological filtering.
FUTURE SCOPE

※The extension of this research would concentrate on the development of


improved algorithms for the rapid and effective identification of diseased
leaves.
※To improve recognition rate in classification process Artificial Neural
Network, Bayes classifier, Fuzzy Logic and hybrid algorithms can also be
used.
THANK YOU

You might also like