10 VI June 2022

https://doi.org/10.22214/ijraset.2022.44226
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue VI June 2022- Available at www.ijraset.com

Plant Disease Detection and Classification Using

Deep Learning
Dr. P. L. Ramteke1, Pooja Wadnere2
1, 2
Computer Science and Engineering, Sant Gadge Baba Amravati University

Abstract: The application of deep learning in plant disease recognition can avoid the disadvantages caused by artificial selection
of disease spot features, make plant disease feature extraction more objective, and improve the research efficiency and
technology transformation speed. In this paper, we present the concept for the detection of plant leaf disease using deep learning
and advanced imaging techniques. We hope that this work will be a valuable resource for researchers who study the detection of
plant diseases and insect pests. At the same time, we also discussed some of the current challenges and problems that need to be
resolved.
Keywords: Deep learning, plant leaf disease detection, visualization, small sample, hyper spectral imaging.

I. INTRODUCTION
The situation of plant diseases has a pessimistic cynical on agricultural production. If plant diseases are not discovered within the
time period, food uncertainty will increase [1]. Early detection is the basis for efficient prevention and control of plant diseases, and
they play a important role in the management and decision making of agricultural production. Now, plant disease detection has been
a critical issue.
India is one of the developing countries wherein majority of population of country is depends on agriculture and agricultural
production. Studies show that the plant leaf disease reduces the quality and quantity of agricultural products. The identification and
recognition of plant leaf disease by open naked eye is quite difficult task for farmers and consult scientist or expertise person is very
costly for farmers in our developing countries like India so the basic motivation behind this dissertation is to detect and identify
disease at early stage is important task for farmers. Detection of disease at early stage can save the whole crops from a disease.

II. LITERATURE SURVEY

There are some research papers previously presented to summarize the research about agriculture (including plant disease
recognition) by DL [2], [3], but they lacked some of the recent developments in terms of visualization techniques implemented
along with the DL and modified the famous DL models, which were used for plant disease identification.
The article [4] presented many imaging techniques for plant disease detection, and the focus was on imaging techniques. The major
techniques presented for plant diseases and classification are SVM, K-means, and KNN.
The article [5] presented many developed/modified DL architectures implemented to detect and classify plant diseases. And
provided a comprehensive explanation of DL models used to visualize various plant diseases. But there is no mention of the early
detection of the diseases and how to detect and classify plant diseases based on small samples.
In the paper [6], the authors had presented a comprehensive review of recent research work done in plant disease recognition using
IPTs, from the perspective of feature extracted based on hand-crafted or using deep learning techniques. And it is concluded that the
deep learning. techniques have superseded shallow classifiers trained using hand-crafted features. But they lacked some of the
recent developments in terms of visualization techniques, and there is no mention of the early detection of the diseases and how to
detect and classify plant diseases based on small samples.
Tian et al. [7] proposed an approach (CycleGAN) that can generate more apple disease images. Generated images augmented by
conditional deep convolutional generative adversarial networks (C-DCGAN) [8] use the segmented tea disease spot image as the
input of VGG16. The result showed that the average accuracy is about 28% higher by using C-DCGAN than rotation and translation.
The article [9] generated images by using deep convolutional generative adversarial networks (DC-GAN), and achieved a top-1
average identification accuracy of 94.33% on GoogLeNet. The T-distribution random neighborhood embedding (T-SNE) verified
that the image distribution generated by this method was closer to the sample distribution of the real image.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 2228
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue VI June 2022- Available at www.ijraset.com

In the paper [10], four different kinds of grape leaf disease images were expanded by a novel Leaf GAN model. The experimental
results showed that the Leaf GAN model could make the grape leaf disease images highlight the disease and generate enough grape
leaf disease images. It was proved that Leaf GAN was superior to those of the DCGAN and WGAN.

III. SYSTEM DESIGN

Fig 1 Traditional image recognition processing

The general process of using traditional image recognition processing technology to identify plant diseases is shown in Fig 1

A. Image Acquisition
Image Acquisition is that the initiative in any image process system. the overall aim of any image acquisition is to transform an
optical image (real-world information) into an array of numerical data that may well be later manipulated on a computer. Image
acquisition is achieved by suitable cameras. We use different cameras for different applications. If we need an X-ray image, we use
a camera (film) that is sensitive to X-rays. If we would like an infrared image, we tend to use cameras that are sensitive to infrared
light. For normal images (family pictures, etc.), we use cameras that are sensitive to the visual spectrum..

B. Image Processing
Based on the number of pixels an image is represented by its dimensions (height and width). For example, if the dimensions of an
image are 500 x 400 (width x height), the total number of pixels in the image is 200000. This pixel is a point on the image that takes
on a specific shade, opacity or colour. It is usually represented in one of the following:
1) Grayscale - A pixel is an integer with a value between 0 to 255 (0 is completely black and 255 is completely white).
2) RGB - A pixel is made up of 3 integers between 0 to 255 (the integers represent the intensity of red, green, and blue).
3) RGBA - It's an extension of RGB with a new alpha field, that represents the opacity of the image.
Image process needs mounted sequences of operations that are performed at every constituent of an image. The image processor
performs the first sequence of operations on the image, pixel by pixel. Once this is often absolutely done, it'll begin to perform the
second operation, and so on. The output value of these operations can be computed at any pixel of the image.
Image process is that the method of transforming an image into a digital kind and activity sure operations to induce some helpful
info from it. The image process system typically treats all pictures as 2d signals once applying sure preset signal process strategies.

There are five main types of image processing:

a) Visualization - Find objects that are not visible in the image
b) Recognition - Distinguish or detect objects in the image
c) Sharpening and Restoration - Create an enhanced image from the original image
d) Pattern Recognition - Measure the various patterns around the objects in the image
e) Retrieval - Browse and search images from a large database of digital images that are similar to the original image

C. Feature Extraction
Feature extraction could be a a part of the dimensionality reduction method, in which, an initial set of the data is split and reduced to
additional manageable teams. therefore after you wish to method it'll be easier. the foremost necessary characteristic of those
massive a large sets is that they need an large number of variables. These variables require a lot of computing resources to process.
So Feature extraction helps to get the best feature from those big data sets by selecting and combining variables into features, thus,
effectively reducing the amount of data. These features are simple to process, however still ready to describe the particular
knowledge set with accuracy and originality.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 2229
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue VI June 2022- Available at www.ijraset.com

The technique of extracting the features is useful when you have a large data set and need to reduce the number of resources without
losing any important or relevant information. Feature extraction helps to reduce the amount of redundant data from the data set. In
the end, the reduction of the data helps to create the model with less machine effort and additionally will increase the speed of
learning and generalization steps within the machine learning method.

D. Identification and Classification

In easy words, image classification could be a technique that's wont to classify or predict the category of a selected object in an
image. the main goal of this system is to accurately establish the features in an image. In general, the image classification
techniques can be categorized as parametric and non-parametric or supervised and unsupervised as well as hard and soft classifiers.
For supervised classification, this technique delivers results based on the decision boundary created, which mostly rely on the input
and output provided while training the model. But, in the case of unsupervised classification, the technique provides the result based
on the analysis of the input dataset own its own; features are not directly fed to the models. the main steps concerned in image
classification techniques are decisive an acceptable arrangement, feature extraction, choosing good training samples, image pre-
processing and choice of acceptable classification technique, post-classification process, and finally assessing the overall accuracy.
during this technique, the inputs are sometimes an image of a selected object, like the rabbit in the above image, and also the outputs
are the expected categories that outline and match the input objects. Convolutional Neural Networks (CNNs) is that the most well
liked neural network model that's used for image classification downside.

IV. RESULT AND DISCUSSION

Graph 1 shows Loss while training the system. The training loss is calculated over the entire training dataset. Train Error involves
the human interpretable metric of your model's performance. Normally it means what percentage of training examples the model got
incorrect. Table 1Model Summary while training system A Training Summary is an aggregate list of all of your professional
development activities during a particular fiscal year. A Training Summary is extremely useful for both staff and supervisors when
conducting an annual performance review. Table 2 shows Confusion Matrix while testing. A confusion matrix, in predictive
analytics, is a two-by-two table that tells us the rate of false positives, false negatives, true positives and true negatives for a test or
predictor. We can make a confusion matrix if we know both the predicted values and the true values for a sample set. Fig. 2 shows
Accuracy of CNN Algorithm Accuracy is the most intuitive performance measure and it is simply a ratio of correctly predicted
observation to the total observations. Graph 2 Training performance graph and table 3 shows model summary of Training Page.
Training data is the data you use to train an algorithm or machine learning model to predict the outcome you design your model to
predict. Fig 3 shows Testing Page. Test data is used to measure the performance, such as accuracy or efficiency, of the algorithm
you are using to train the machine. Fig 4 shows Output Module. This page can choose image predict the disease.

Graph 1 Loss while training the system

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 2230
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue VI June 2022- Available at www.ijraset.com

Table 1 Model Summary while training system

Table 2: Confusion Matrix while testing

Fig 2: Accuracy of CNN Algorithm

Graph 2: Training performance

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 2231
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue VI June 2022- Available at www.ijraset.com

Table 3: Model Summary

Fig 3: Testing Page

Fig 4: Output Module

V. CONCLUSION
The implementation of this dissertation is done in plant leaf disease recognition using deep learning. Provided sufficient data is
available for training, deep learning techniques are capable of recognizing plant leaf diseases with high accuracy. The importance of
collecting large datasets with high variability, data augmentation, transfer learning, and visualization of CNN activation maps in
improving classification accuracy, and the importance of small sample plant leaf disease detection and the importance of hyper-
spectral imaging for early detection of plant disease have been discussed. In most of the researches, the Plant Village dataset was
used to evaluate the performance of the DL models. Although this dataset has a lot of images of several plant species with their
diseases, it was taken in the lab. Therefore, in future it is expected to establish a large dataset of plant diseases in real conditions.

REFERENCES
[1] F. Fina, P. Birch, R. Young, J. Obu, B. Faithpraise, and C. Chatwin, ‘‘Automatic plant pest detection and recognition using k-means clustering algorithm and
correspondence filters,’’ Int. J. Adv. Biotechnol. Res., vol. 4, no. 2, pp. 189–199, Jul. 2013.
[2] A. Kamilaris and F. X. Prenafeta-Boldú, ‘‘Deep learning in agriculture: A survey,’’ Comput. Electron. Agricult., vol. 147, pp. 70–90, Apr. 2018.
[3] K. Singh, B. Ganapathysubramanian, S. Sarkar, and A. Singh, ‘‘Deep learning for plant stress phenotyping: Trends and future perspectives,’’ Trends Plant Sci.,
vol. 23, no. 10, pp. 883–898, Oct. 2018.
[4] V. Singh, N. Sharma, and S. Singh, ‘‘A review of imaging techniques for plant disease detection,’’ Artif. Intell. Agricult., vol. 4, pp. 229–242, Oct. 2020.
[5] M. H. Saleem, J. Potgieter, and K. M. Arif, ‘‘Plant disease detection and classification by deep learning,’’ Plants, vol. 8, no. 11, pp. 468–489, Oct. 2019.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 2232
 International Journal for Research in Applied Science & Engineering Technology (IJRASET)
ISSN: 2321-9653; IC Value: 45.98; SJ Impact Factor: 7.538
Volume 10 Issue VI June 2022- Available at www.ijraset.com

[6] L. C. Ngugi, M. Abelwahab, and M. Abo-Zahhad, ‘‘Recent advances in image processing techniques for automated leaf pest and disease recognition—A
review,’’ Inf. Process. Agricult., vol. 180, pp. 26–50, Apr. 2020.
[7] Y. Tian, G. Yang, Z. Wang, E. Li, and Z. Liang, ‘‘Detection of apple lesions in orchards based on deep learning methods of CycleGAN and YOLOV3-
dense,’’ J. Sensors, vol. 2019, pp. 1–13, Apr. 2019.
[8] H.-Y. Wu, ‘‘Identification of tea leaf’s diseases in natural scene images based on low shot learning,’’ M.S. thesis, Dept. Inf. Eng., Anhui Univ., Hefei, China,
2020.
[9] Q. Wu, Y. Chen, and J. Meng, ‘‘DCGAN-based data augmentation for tomato leaf disease identification,’’ IEEE Access, vol. 8, pp. 98716–98728, 2020.
[10] B. Liu, C. Tan, S. Li, J. He, and H. Wang, ‘‘A data augmentation method based on generative adversarial networks for grape leaf disease identification,’’
IEEE Access, vol. 8, pp. 102188–102198, 2020.

©IJRASET: All Rights are Reserved | SJ Impact Factor 7.538 | ISRA Journal Impact Factor 7.894 | 2233