Plant Leaf Disease Detection using CNN

D.Iruthaya Antony Prethika, P.G, Student,

Department of Computer Science and Engineering, Thoothukudi
V. Revathy, Assistant Professor,
Department of Computer Science and Engineering, Thoothukudi
Address
[email protected]
Abstract— Plants and crops that are infected by pests diseases as per that provide medical treatment. Proposed Deep
have an impact on the country's agricultural CNN model has been compared with popular transfer learning
production. Usually, farmers or professionals keep a approach such as VGG16. Plant leaf disease detection has
close eye on the plants in order to discover and wide range of applications available in various fields such as
Biological Research and in Agriculture Institute. Plant leaf
identify diseases. However, this procedure is
disease detection is the one of the required research topic as it
frequently time-consuming costly, andimprecise. may prove benefits in monitoring large fields of crops, and
Plant disease detection can be done by looking for a thus automatically detect the symptoms ofdiseases as soon as
spot on the diseased plant's leaves. The goal of this they appear on plant leaves.
paper is to create a Disease Recognition Model that is
supported by leaf image classification. To detect
plant diseases, we are utilizing image processing with II. DATA COLLECTION AND PROCESSING
a Convolution neural network (CNN). A To detect plant leaf diseases, the first step is to collect a
convolutional neural network(CNN) is a form of dataset of images with both healthy and diseased leaves that
artificial neural network that is specifically intended accurately represent real-world scenarios. The collected
to process pixel input and is used in image images are preprocessed by removing noise and irrelevant
information through techniques such as normalization and
recognition.Farmers do not expertise in leaf disease
data augmentation.
so they produce less production. Plant leaf diseases analytics we follow the below steps :
detection is the important because profit and loss are
depends on production. 1. Data Collection
2. Preprocessing
Index Terms: Image processing,Crops,Support 3. Model selection
vector Machine,Plant disease,Classification. 4. Data Visualization
5. Data Modeling
6. Data Training
Keywords— Heart Disease, Diabetes, Machine Learning.
7. Data Testing
8. Data Prediction
I. INTRODUCTION 9. Evaluation
Agricultural production is a very old means of obtaining 10. Optimization.
food. It is a vital source of income for people all around the
world. No one can exist in our world without food. Plants are
crucial not only for humans, but also for animals who rely on
them for food, oxygen, and other necessities. The government
and experts are takingsignificant initiatives to enhance food
production, and they are working successfully in the real
world.environment, Plants and crops that are infected by pests
have an impact on the country's agricultural
production.Usually, farmers or professionals keep a close eye
on the plants in order to discover and identify
diseases.imprecise. Plant disease detection can be done by
looking for a spot on the diseased plant's leaves. The goal of
this paper is to create a Disease Recognition Model that is
supported by leaf image classification. To detect plant
diseases, we are utilizing image processing with a
Convolution neural network (CNN). A convolutional neural
network(CNN) is a form of artificial neural network that is
specifically intended to process pixel input and is used in
image recognition.Farmers do not expertise in leaf disease
so they produce less production. Plant leaf diseases detection 1.The first step is to collect data. We are using the
is the important because profit and loss are depends on PlantVillage Dataset, which is widely available. This dataset
production. CNN is the solution for leaf disease detection and was released by crowdAI.
classification. Main aim of this research is to detect the apple, 2.Pre-processing and Augmentation of the collected
grape, corn, potato and tomato plants leaf diseases. Plant leaf dataset is done using pre-processing and Image-data
diseases are monitoring of large fields of crops disease generator API by Keras.
detection, and thus automatically detected the some feature of
3.Building CNN(Convolutional Neural In unsupervised learning we have less information about
Network) Model (Vgg-19 architecture) for objects, in particular, the train set is unlabeled. What is our
classification of various plant diseases. goal now? It’s possible to observe some similarities between
4.Developed model will be deployed on the Android groups of objects and include them in appropriate clusters.
Application with help of TensorFlow Some objects can differ hugely from all clusters, in this way
we assume these objects to be anomalies.
III. DESIGN IMPLEMENTATION AND TESTING

Image processing algorithms, particularly deep convolutional

neural networks (CNNs), have gained popularity as an
effective approach to detect and diagnose diseases in plants.
These algorithms analyze leaf images and use machine
learning techniques to identify and classify diseases
accurately. 3)Semi-Supervised learning
Semi-supervised learning tasks include both problems we
described earlier: they use labeled and unlabeled data. That is
a great opportunity for those who can’t afford labeling their
A. Module Building data. The method allows us to significantly improve accuracy,
There is a very wide range of machine learning because we can use unlabeled data in the train set with a small
algorithms to choose from, most of which are available in the amount of labeled data.
python library Scikit-learn. However, most of the
implementations of these algorithms do not accept sparse
matrices as inputs, and since we have a large number of
nominal features coming from our n-grams features it is
imperative that we encode our features in a sparse matrix. Out
of the algorithms that do support sparse matrices in Scikit-
learn, I ended up trying naive Bayes, logistic regression and
support vector machine (SVM) with a linear kernel. I got the
best results in cross validation using SVM with aneuclidean
regularization coefficient of 0.1. 4)Reinforcement learning
 Supervised learning
Reinforcement learning is not like any of our previous
tasks because we don’t have labeled or unlabeled datasets
 Unsupervised learning
here. RL is an area of machine learning concerned with how
software agents ought to take actions in some environment to
 Semi Supervised learning
maximize some notion of cumulative reward.
 Reinforcement
1.Data Collection
.
Data collection is the gathering of information from
1)Supervised learning
multiple sources, and data analytics is the
Supervised learning is the task of inferring a function from processing of that information in order to derive
labeled training data. By fitting to the labeled training set, we usable insights. To process and gain insights from
want to find the most optimal model parameters to predict data acquired from various sources and methods,
unknown labels on other objects (test set). If the label is a real specific data analysis methodologies and tools are
number, we call the task regression. If the label is from the required.
limited number of values, where these values are unordered,
then it’s classification. Training Dataset:
To detect plant leaf diseases, the first step is to collect a
dataset of images with both healthy and diseased leaves
that accurately represent real-world scenarios. The
collected images are preprocessed by removing noise
and irrelevant information through techniques such as
normalization and data augmentation.The dataset used
consisted of over 20,000 images, and CNN was used
for classification, including 12. classes for diseased
2)UnSupervised learning leaves and 3 classes for healthy leaves. The model
achieved an accuracy of 98.29% for. training and
98.029% for the testing dataset
The data's in the file are:

3.Implementation work:

Apple, grape, potato, and tomato plant leaves which are

categorized total 24 types of labels apple label namely: Apple
scab, Black rot, Apple rust, and healthy. Corn label namely:
Corn Cercospora Gray spot, Corn rust, Corn healthy, Corn
Blight [11][13]. Grape label namely: Black rot, Esca, healthy,
and Leaf blight. Potato label namely: Early blight, healthy,
Testing Dataset: and Late blight. Tomato label namely: bacterial spot, early
blight, healthy, late blight, leaf mold, septoria leaf spot, spider
To detect plant leaf diseases, the first step is to collect a mite, target sport, mosaic virus, and yellow leaf curl virus[11]
dataset of images with both healthy and diseased leaves that [13].The dataset consist of 31,119 images of apple, corn,
accurately represent real-world scenarios The collected grape, potato and tomato, out of 31,119 images 24000 images
images are preprocessed by removing noise and irrelevant are used. all Images are resized into 256 x 256,that images
information through techniques such as normalization and divided into two parts training and testing dataset, the whole
data augmentation. range of the train test split using 80-20 (80% of the
2.Disease leaf Dataset:

The first step is to perform data cleaning. If there’s any

incorrect, redundant, irrelevant or partially formatted data that
will be removed or changed, as these data are not beneficial
nor necessary because it might produce false results. Data
cleaning is the process of eliminating or changing data that is
erroneous, incomplete, irrelevant, redundant, or incorrectly
formatted in order to prepare it for analysis. When it comes to
data analysis, this data is usually not necessary or beneficial
because it can slow down the process or produce false results.
To reduce this we first find months for the dates given in the
dataset. Then the wind speed is rewritten for 0.5 intervals. For
example: wind speeds between 3.25 and 3.75 turns 3.5,wind
speeds between 3.75 and 4.25 turns 4.0.The wind direction is
rewritten for 30 intervals. For example: wind directions
between 15 and 45 turns 30, wind speeds between 45 and 75
turns 60 Implementation

whole dataset used for the training and 20% for the testing)
[11][13]. Then train CNN model

4.Data Visualization
The graphical depiction of information and data is known as The project demonstrated the effectiveness of deep learning
data visualization. Data visualization tools make it easy to techniques in detecting plant leaf diseases.The project
examine and comprehend trends, outliers, and patterns in data
by employing visual elements like charts, graphs, and maps.
Here we data analysis because it helps us to spot developing
patterns

Training Vs Test model Model

achieved high accuracy rates in classifying the different types

of plant leaf diseases using a Convolutional Neural Network
(CNN) architecture.
A 95.6% accuracy rate was achieved using early stopping
while Training the model on 50 epochs. Figure 7 depicts the
visualization of training and validation accuracy. The result
of detecting and recognizing a strawberry plant is shown in
Figure 8. On the left, a healthy plant leaf, and on the right, a
sick infected plant. The result of detecting and recognizing a
potato plant is shown in. On the left, a healthy plant leaf,
and on the right, a sick infected plant.

V.REFERENCES
Proposed workflow
Convolutional neural networks (CNN) can be used for the [1]. Liu, Bin, "Identification of apple leaf diseases based on
computational model creation that works on the deep convolutional neural networks
unstructured image inputs and converts to output labels of
corresponding classification.complexity of the extracted [2]. Jeon, Wang-Su, and Sang-Yong Rhee. "Plant leaf
features increases. The size of the filter is fixed to 5 × 5 recognition using a convolution neural network."
whereas number of filters is increased progressively as we International Journal of Fuzzy Logic and Intelligent Systems
move from one block to another. The number of filters is 17.1 (2017): 26-34.
20 in the first convolution block while it is increased to [3]. Amara, Jihen, Bassem Bouaziz, and Alsayed
50 in the second and 80 in the third.. Algergawy. "A Deep Learning-based Approach for Banana
Leaf Diseases Classification." BTW (Workshops). 2017.

Figure 4. Work flow [4]. Lee, Sue Han, et al. "How deep learning extracts and
learns leaf features for plant classification." Pattern
Recognition 71 (2017): 1-13.
[5]. Sladojevic, Srdjan, et al. "Deep neural networks-based
recognition of plant diseases by leaf image classification."
IV.CONCLUSIONS Computational intelligence and neuroscience 2016 (2016).
 [6]. Lee, Sue Han, et al. "Plant Identification System based and Applied Mathematics Special Issue 882 Disease
on a Convolutional Neural Network for the LifeClef 2016 detection using Image Processing Techniques,” IOSR J. of
Plant Classification Task." CLEF (Working Notes). 2016. Compu. Eng
[7]. He, Kaiming, et al. "Deep residual learning for image
[10]. Y. Q. Xia, Y. Li, and C. Li, “Intelligent Diagnose
recognition." Proceedings of the IEEE conference.
System of Wheat Diseases Based on Android Phone,” J. of
[8].K.Padmavathi, and K.Thangadurai,“Implementation of Infor. & Compu. Sci., vol. 12, pp. 6845-6852, Dec. 2015.
RGB and Gray scale images in plant leaves disease
detection –comparative study,”
[9]. Kiran R. Gavhale, and U. Gawande, “An Overview of
the Research on Plant Leaves International Journal of Pure