Global Transitions Proceedings 3 (2022) 305–310

Contents lists available at ScienceDirect

Global Transitions Proceedings

journal homepage: http://www.keaipublishing.com/en/journals/global-transitions-proceedings/

Plant leaf disease detection using computer vision and machine learning
algorithms
Sunil S. Harakannanavar a,∗, Jayashri M. Rudagi b, Veena I Puranikmath b, Ayesha Siddiqua a,
R Pramodhini a
a
Department of Electronics and Communication Engineering, Nitte Meenakshi Institute of Technology, Yelahanka, Bangalore, Karnataka, India
b
Department of Electronics and Communication Engineering, S. G. Balekundri Institute of Technology, Shivabasava Nagar, Belagavi, Karnataka, India

a r t i c l e i n f o a b s t r a c t

Keywords: Agriculture provides food to all the human beings even in case of rapid increase in the population. It is recom-
Convolutional Neural Networks mended to predict the plant diseases at their early stage in the field of agriculture is essential to cater the food to
Discrete Wavelet Transform the overall population. But it unfortunate to predict the diseases at the early stage of the crops. The idea behind
Principal Component Analysis
the paper is to bring awareness amongst the farmers about the cutting-edge technologies to reduces diseases in
Nearest Neighbor
plant leaf. Since tomato is merely available vegetable, the approaches of machine learning and image processing
Leaf disease
with an accurate algorithm is identified to detect the leaf diseases in the tomato plant. In this investigation, the
samples of tomato leaves having disorders are considered. With these disorder samples of tomato leaves, the farm-
ers will easily find the diseases based on the early symptoms. Firstly, the samples of tomato leaves are resized to
256 × 256 pixels and then Histogram Equalization is used to improve the quality of tomato samples. The K-means
clustering is introduced for partitioning of dataspace into Voronoi cells. The boundary of leaf samples is extracted
using contour tracing. The multiple descriptors viz., Discrete Wavelet Transform, Principal Component Analysis
and Grey Level Co-occurrence Matrix are used to extract the informative features of the leaf samples. Finally,
the extracted features are classified using machine learning approaches such as Support Vector Machine (SVM),
Convolutional Neural Network (CNN) and K-Nearest Neighbor (K-NN). The accuracy of the proposed model is
tested using SVM (88%), K-NN (97%) and CNN (99.6%) on tomato disordered samples.

1. Introduction The key issues and challenges [10,11] are identified by the re-
searchers and the scientists, while analysing the leaf diseases of plant.
Developed Technologies have provided the ability to produce suffi- Some of them are as follows
cient food to meet the demand of society. But still, the safety and se-
curity of the food or crops remained unattained. Factors like change in 1 The quality of the leaf image must be high.
climate, the decline in pollinators, Plant disease, and others are chal- 2 Publicly available Dataset requirement.
lenging to the farmers. An important foundation for these factors needs 3 Noisy data affecting the leaf samples.
to be attained on a priority basis [1,2]. Making use of analysis and de- 4 Through the process of segmentation, diseases may be identified but
tection processes using present technology helps the farmers to get rid the samples must undergo training and testing.
of such problems. During pandemic situations like COVID 19 the nation 5 Classification is one more challenge, in the stage of detecting the leaf
is dependent on the recent technologies to prevent address the issues to diseases.
reduce the transmission of the diseases [2–5]. As plant diseases are a sig- 6 Color of the leaves may be varied due to environmental effect.
nificant threat to human life as they may lead to droughts and famines. 7 Variety of diseases can be seen in various kinds of plants, so detection
In turn it results causing substantial losses, where farming is accompa- of disease is quite difficult.
nying in commercial purpose. The use of technologies like Computer
vision and Machine Learning (ML) helps to fight against diseases [6–8]. Based on the challenges discussed above and combined techniques
In this paper, we are using ML to give a solution to Plant Diseases. In this using image processing (IP) and ML, the proposed model provide better
method, we have divided the process into three stages Identity, Analyse accuracy. Keeping all these things in mind, in this paper an algorithm
and Verify with the Available database [9]. based on ML and IP tools to automatically detect leaf diseases is pro-
posed.

∗
Corresponding author.
E-mail address: sunilsh143@gmail.com (S.S. Harakannanavar).

https://doi.org/10.1016/j.gltp.2022.03.016

Available online 2 April 2022

2666-285X/© 2022 The Authors. Publishing Services by Elsevier B.V. on behalf of KeAi Communications Co. Ltd. This is an open access article under the CC BY
license (http://creativecommons.org/licenses/by/4.0/)
S.S. Harakannanavar, J.M. Rudagi, V.I. Puranikmath et al. Global Transitions Proceedings 3 (2022) 305–310

The contribution for the above proposed framework is done in three

stages. Firstly, the HE and K-means clustering are employed to maxi-
mize the quality and segment the leaf samples. Based on the K-means
clustering response, the leaf is diseased or not can be predicted at the
early stage of operation. Secondly, The DWT, PCA and GLCM are used
to extract the informative regions/features of the samples. Lastly, as a
part of machine learning approaches the SVM, KNN and CNN are used
to classify the features
Section 2 describes the detailed description of existing leaf disease
techniques. The leaf disease detection model is discussed in section 3.
The results of proposed model on tomato leaf samples having six disor-
ders are analysed in section 4. The leaf detection model is concluded in
Section 5.

2. Literature survey

Lot of work has been devoted to the detection of leaf diseases using
image processing in the history and it continues to attract research to
carry out their research work in this field. Automatic crop disease de-
tection using image processing and machine learning has been gaining
prominence in recent years.
P. Krithika et al., [9] pre-processed by image resizing, contrast en-
hancement and color-space conversion. The K-Means clustering for seg-
mentation and feature extraction using GLCM is performed. Classifica-
tion was made using multiclass SVM. R. Meena et al., [10] performed
color space conversion followed by enhancement process. The primary
colors of leaves are converted into L∗ A∗ B∗ . The K-Mean clustering algo-
rithm is used for segmentation. The GLCM and SVM are used for feature
extraction and classification respectively. Bharat et al., [12] acquired
images using digital camera and median filter is used for image en-
hancement. K-Mean clustering is used for segmentation. SVM is used
for classification. Pooja et al., [13] segmentation is done to get the ar-
eas of interest that is the infected region. It is done using k-Mean clus-
tering algorithm, Otsu’s detection converting RGB to HSI later segmen-
tation is done using boundary and spot detection algorithm. Rukaiyya
et al., [14] performed pre-processing by contrast adjustment and nor-
Fig 1. Proposed model.
malization. The conversion of color transform into YCBCR and Bi-level
thresholding is performed. The GLCM, and HMM are used for features
extraction and classification [15].
Chaitali et al., [16] segmentation of image is applied for background of the leaf disease was carried using k-means clustering and the classifi-
subtraction. The classification approach is carried out by KNN, ANN cation of the disease was performed using SVM. Statistical information
and SVM method. In KNN, it classifies samples using nearest distance such as mean, median, mode, standard deviation was used by authors
between trained and testing subjects [17]. Varun et al., [19] has devel- to record their findings. Vidyashree Kanbur et al., [24] developed the
oped model for extraction thresholding technique and morphological model for leaf detection disease using multiple descriptors. The model
operation. Then multiclass SVM is used as classifier. For segmentation, was tested on local leaf database and the performance of the model was
based on a set of marks generated by analysis of the color and lumi- superior., but it can be tested on publicly available dataset.
nosity components of different regions of image is L∗ A∗ B∗ color spaces. Pushpa et al., [25] Indices Based Histogram technique is used to seg-
The GLCM is used for feature extraction. Vijai Singh et al., [19] consid- ment unhealthy region of the leaf. The authors have surpassed other
ered samples of plant leaves like rose/beans (bacterial disorder), lemon segmentation techniques such as slice segmentation, polygon approx-
(sun burn disorder), banana (early scorch) and beans (fungal) that are imation, and mean-shift segmentation. Kaleem et al., [26] considered
captured using a digital camera. The green regions as background us- pre-processed to resize them into 300∗ 300 sized images, remove back-
ing thresholding algorithm. Finally, the genetic algorithm is used to get ground noise, enhance brightness, and adjust the contrast. The K-means
the segmented image. The color co-occurrence is adapted for useful ex- clustering for segmentation and the useful features are extracted using
traction of features from the segmented images. The Minimum Distance Statistical GLCM and SVM classifier is used for classification of leaf dis-
Criterion and then SVM classifier is used for classification purpose. The orders.
average accuracy of 97.6% has been recorded.
Sa’ed Abed et al., [20] performed scaling and stretching (min-max 3. Proposed model
linear) process for the input samples to improve the quality. The creation
of HIS model is completed and the same is segmented later. The tech- The model is developed based on the IP and ML approaches for de-
niques of combined Euclidean distance and K-mean clustering is per- tection of leaf disease in presented in this section. The proposed model
formed for segmentation of the samples. The GLCM and SVM are used (DWT+PCA+GLCM+CNN) using computer vision and machine learning
for feature extraction and classification respectively. Arya et al., [21,22] approaches for leaf disease detection is shown in Fig. 1.
takes input RGB image and creates color transformation then conversion The tomato samples having six disorders are considered to evaluate
of the input samples to HIS format. Finally, segment the components us- its accuracy and to recognize the leaf disease as Healthy or Unhealthy.
ing Otsu’s method. Nema et al., [23] images of 81 were included in the As a part of image processing, the samples of tomato are resized to
database and analysis was performed in L∗ a∗ b color space. Segmentation 256 × 256 pixels to maintain equal in their size throughout the exper-

306
S.S. Harakannanavar, J.M. Rudagi, V.I. Puranikmath et al. Global Transitions Proceedings 3 (2022) 305–310

Fig 2. Sample of tomato having disorders.

Fig 4. Contour tracing.

Fig 3. K-mean clustering.

Fig 5. DWT decomposition.

iment. The HE and K-means clustering are employed to maximize the

quality and segment the leaf samples. Based on the K-means clustering 3.3. Feature extraction
response, the leaf is diseased or not can be predicted at the early stage
of operation. The boundaries of the leaf samples can be extracted using Discrete Wavelet Transform: The DWT [23] is applied on enhanced
contour tracing. tomato samples to extract the useful features. The DWT decomposes into
The DWT, PCA and GLCM are used to extract the informative re- sub-bands of lower (LL, LH, HL) and higher frequency (HH) components.
gions/features of the samples. In the next stage as a part of machine The LL component of DWT carries maximum availability of information
learning approaches the SVM, KNN and CNN are used to classify the when compared with higher frequency components of DWT as shown
features and the performance of the model is recorded. in Fig. 5.
Gray Level Co-occurrence Matrix: The optimal features are selected
obtained from wavelet decomposition is carried out by Principal Com-
3.1. Dataset ponent Analysis [24,25]. The GLCM uses in the distribution of higher
order of gray values that are defined with neighborhood criterion
The village database of tomato leaf [15,34] is considered, the plants [26,27]. The several properties are derived from the GLCM technique
which are affected from variety of diseases. The images of tomato leaf for extraction of leaf features.
having six disorders are taken to carry out the experiments for detection The most used texture-based features are as follows. Homogeneity,
of leaf disease. The samples of leaf images in the database are shown in Autocorrelation, Dissimilarity, Entropy, Sum of squares, Average, Vari-
Fig, 2. ance, Entropy [28,29]. The features obtained using DWT, GLCM and
PCA are combined to form feature vector which are provided as an in-
put sample to the classifiers to recognize classify the images.
3.2. Preprocessing

K-means clustering technique [17–21] is applied on leaf images to 3.4. Classification

find out the infected region. The K-mean clustering is used to get the data
centre of the image and make the clusters of that image and calculates The techniques such as SVM, KNN and CNN are used for classify-
the centre distance from the other cluster. Samples of leaf after applying ing the samples. The CNN [30–32] is a type of ANN which is designed
k-mean clustering algorithm [20] is shown in Fig. 3. to process the data. The architecture of CNN includes input (IL), output
Contour tracing [21,22]is performed on digital leaf samples to ex- (OL) and hidden layers (HL), which are multiple in its nature. The HL in-
tract their general shape information. After extracting the contour, its cludes convolutional layers, RELU layer i.e., which performs activation
characteristics is analysed and used for pattern classification. It of- function, pooling, fully connected and normalization layers [33,34]. It
ten helps for determining the efficiency of feature extraction process is having mathematically evident that its architecture is cross correla-
[19,20]. The images appeared after performing contour tracing is shown tion rather than a convolution and demonstrates significance for the
in Fig. 4. indices in the matrix [35]. Regular 3-layer network is shown in Fig. 6.

307
S.S. Harakannanavar, J.M. Rudagi, V.I. Puranikmath et al. Global Transitions Proceedings 3 (2022) 305–310

Fig 6. A regular 3-layer Neural Network.

Fig 8. Performance of CNN Classifier.

Fig 7. CNN Confusion matrix.

The confusion matrix [36,37] for CNN having output class and tar-
get class is shown in Fig. 7. The progress of training samples of leaf fea-
tures classified using CNN to know the accuracy and errors are shown
in Fig. 8.

3.5. Evaluation of Leaf disease

The parameters such as Precession, Recall and F-measure [38] for

the proposed model is calculated and is given in Eqs. 1,2, and 3. Fig 9. Comparison of Proposed model.
True Positives
Pr ecision Measur e (%) = X100 (1)
True Positives + False Positives
The model is validated by training/testing the dataset samples. The
True Positives hardware and software specifications are recommended to carry out the
Recall Measure (%) = X100 (2)
True Positives + False Negatives work. ie., operation system: Windows10, Core: GPU-NVIDIA, language:
2∗ Precision ∗ Recall Python, Libraries: Image data generator, Open Cv, Kera’s, tensor flow,
F − measure = X100 (3) NumPy, Mat plot, Dataset: Plant village dataset having Tomato samples
Precision + Recall
(Six disorder). The parameters such as Precision, Recall and F1 score are
4. Experimental results used to evaluate the performance of proposed model.
The proposed model is tested on tomato leaf disease dataset with an
The samples of tomato leaf of village dataset is considered to carry overall 600 samples. Fig. 9 shows the comparison of proposed model
out to evaluate the proposed model. The 100 samples of healthy leaf with the existing models. The precession, recall and F1-scores of SVM,
is taken to test on the proposed model. The model results in idetifying KNN and CNN is tabulated in Table 1.
99 samples with an accuracy of 99%. In the 100 Mosaic virus tomato The result of proposed model is compared with existing
samples, the model recognizes 100 samples with an accuracy of 100%. models. It is observed that, the accuracy of Proposed model
For leaf mold category, the model results in the accuracy of 100%. For (DWT+PCA+GLCM+CNN) provides better accuracy of 99.09% com-
the 100 samples of yellow curl, the model performed 99%. Similarly, pared to the other existing models.
the Spotted spider mite and Target Spot results in 99% and 100% re- The proposed model (DWT+PCA+GLCM+CNN) using computer vi-
spectively. Overall 600 samples of tomato village dataset is tested on sion and ML classification technique is compared with the methodology
the proposed model for evaluation, as a result, the model provides an explained by Hossain et al., [22] and Vidyashree et al., [24] and Thanjai
better accurcay of 99.5%. Vadivel et al., [34] and is tabulated in Table 2. The accuracy obtained

308
S.S. Harakannanavar, J.M. Rudagi, V.I. Puranikmath et al. Global Transitions Proceedings 3 (2022) 305–310

Table 1
Comparison of different classification techniques.

Techniques Disorders Precision Recall F1-Score Accuracy (%)

K mean clustering + GLCM + SVM Healthy 0.983 0.834 0.881 81.9

MV 0.913 0.722 0.821 86.8
LM 0.895 0.817 0.896 80.2
YC 0.811 0.889 0.992 86.3
SM 0.977 0.861 0.985 85.9
TS 0.987 0.898 0.994 87.9
Overall 0.967 0.978 0.915 89
LBP+PCA+K mean clustering+ KNN Healthy 0.858 0.998 0.905 95.9
MV 0.834 0.852 0.932 92.8
LM 0.722 0. 844 0.916 89.6
YC 0.817 0.985 0.892 87.7
SM 0.889 0.876 0.915 88.9
TS 0.861 0.928 0.958 94.8
Overall 0.978 0.975 0.952 97.3
DWT+PCA+GLCM+CNN(Proposed method) Healthy 0. 985 0.984 0.979 98.9
MV 0.949 0.945 0.962 95.3
LM 0.932 0. 934 0.956 95.5
YC 0.938 0.968 0.979 96.4
SM 0.953 0.919 0.899 97.2
TS 0.967 0.978 0.898 95.5
Overall 0.995 0.995 0.988 99.09

Table 2
Comparison of existing methodologies with proposed model.

Authors Methodologies/Descriptions Accuracy (%)

Hossain et al., [22] Based on univariate statistical features test + SVM 90

Vidyashreeet al., [24] K-mean clustering + GLCM + SVM 90
Thanjai Vadivel et al., [34] Fast Enhanced Learning Method 99
Proposed Model DWT+PCA+GLCM+CNN 99.09

by the proposed method is better compared with the existing method- [8] K. Yu, L. Lin, M. Alazab, L. Tan, B. Gu, Deep learning-based traffic safety solution
ologies. for a mixture of autonomous and manual vehicles in a 5G-enabled intelligent trans-
portation system, IEEE Trans. Intell. Transp. Syst. 22 (7) (2020) 4337–4347.
[9] Dataset Kaggle, https://www.kaggle.com/thanjaivadivelm
5. Conclusion and future scope [10] Chaitali G Dhaware, K H Wanjale, A modern approach for plant leaf disease classi-
fication which depends on leaf image processing, in: IEEE International Conference
The proposed model uses computer vision techniques including RGB on Computer Communication and Informatics, 2017, pp. 12–16.
[11] J.P. Nayak, K. Anitha, B.D. Parameshachari, R. Banu, P. Rashmi, PCB Fault detection
conversion to gray, HE, K-means clustering, contour tracing is employed using Image processing, IOP Conference Series: Materials Science and Engineering,
in preprocessing stage. The multiple descriptors Discrete Wavelet Trans- 225, IOP Publishing, 2017.
form, Principal Component Analysis and GLCM are used to extract [12] R P Narmadha, G Arulvadivu, Detection and measurement of paddy leaf disease
symptoms using image processing, in: IEEE International Conference on Computer
the informative features of the leaf samples. The machine learning ap- Communication and Informatics, 2017, pp. 1–4.
proaches such as SVM, K-NN and CNN are used to distinguish diseased [13] C.H. Nguyen, T.L. Pham, T.N. Nguyen, C.H. Ho, T.A. Nguyen, The linguistic sum-
or non-diseased leaf. The analysis of the proposed model is well suited marization and the interpretability, scalability of fuzzy representations of multi-
level semantic structures of word-domains, Microprocessors Microsyst. 81 (2021)
for CNN machine learning classification technique with a desired accu- 103641.
racy compared to other state of the art method. In future, the model [14] V Gupta, N Sengar, M Dutta, C Travieso, J Alonso, Automated segmentation of pow-
can be improved using fusion techniques for extraction of significant dery mildew disease from cherry leaves using image processing, in: IEEE Interna-
tional Conference and Workshop on Bioinspired Intelligence, 2017, pp. 1–4.
features and examined for other leaf samples of datasets. [15] K. Yu, L. Tan, L. Lin, X. Cheng, Z. Yi, T. Sato, in: Deep-learning-empowered breast
cancer auxiliary diagnosis for 5GB remote E-health, 28, IEEE Wireless Communica-
References tions, 2021, pp. 54–61.
[16] Vijai Singh, A.K. Misra, Detection of plant leaf diseases using image segmentation
[1] P Krithika, S Veni, Leaf disease detection on cucumber leaves using multiclass sup- and soft computing techniques, Inf. Process. Agricult. 4 (1) (2017) 41–49.
port vector machine, in: IEEE International Conference on Wireless Communica- [17] S Abed, A Esmaeel, A novel approach to classify and detect bean diseases based
tions, Signal Processing and Networking, 2017, pp. 1276–1281. on image processing, in: IEEE Symposium on Computer Applications & Industrial
[2] R Prakash, G P Saraswathy, G Ramalakshmi, Detection of leaf diseases and classifica- Electronics, 2018, pp. 297–302.
tion using digital image processing, in: IEEE International Conference on Innovations [18] S.M. Nagashetti, S. Biradar, S.D. Dambal, C.G. Raghavendra, B.D. Parameshachari,
in Information, Embedded and Communication Systems, 2017, pp. 1–4. Detection of disease in Bombyx Mori Silkworm by using image analysis approach, in:
[3] B Mishra, S Nema, M Lambert, S Nema, Recent technologies of leaf disease detec- 2021 IEEE Mysore Sub Section International Conference (MysuruCon), IEEE, 2021,
tion using image processing approach-review, in: IEEE International Conference on pp. 440–444.
Innovations in Information, Embedded and Communication Systems, 2017, pp. 1–5. [19] M Arya, K Anjali, D Unni, Detection of unhealthy plant leaves Using image processing
[4] C. Puttamadappa, B.D. Parameshachari, Demand side management of small scale and genetic algorithm with Arduino, in: IEEE International Conference on Power,
loads in a smart grid using glow-worm swarm optimization technique, Microproces- Signals, Control and Computation, 2018, pp. 1–5.
sors Microsyst. 71 (2019) 102886. [20] G. Chandrasekaran, T.N. Nguyen, D Hemanth, Multimodal sentimental analysis for
[5] V Pooja, R Das, V Kanchana, Identification of plant leaf diseases using image pro- social media applications: A comprehensive review, Wiley Interdisciplinary Rev. 11
cessing techniques, in: IEEE International Conference on Technological Innovations (5) (2021) e1415.
in ICT for Agriculture and Rural Development, 2017, pp. 130–133. [21] S Hossain, R Mou, M Hasan, S Chakraborty, A Razzak, Recognition and detection of
[6] D.L. Vu, T.K. Nguyen, T.V. Nguyen, T.N. Nguyen, F. Massacci, P.H. Phung, HIT4Mal: tea leaf’s diseases using support vector machine, IEEE Int. Colloquium Signal Process.
hybrid image transformation for malware classification, Trans. Emerging Telecom- Appl. (2018) 150–154.
mun. Technol. 31 (11) (2020) e3789. [22] J. Zhang, K. Yu, Z. Wen, X. Qi, A.K. Paul, 3D reconstruction for motion blurred im-
[7] R P Shaikh, S A Dhole, Citrus Leaf Unhealthy Region Detection by using Image Pro- ages using deep learning-based intelligent systems, CMC-Comput. Mater. Continua
cessing Technique, in: IEEE International Conference on Electronics, Communication 66 (2) (2021) 2087–2104.
and Aerospace Technology, 2017, pp. 420–423.

309
S.S. Harakannanavar, J.M. Rudagi, V.I. Puranikmath et al. Global Transitions Proceedings 3 (2022) 305–310

[23] S. Nema, A. Dixit, Wheat Leaf Detection and Prevention Using Support Vector Ma- [31] MH Saleem, S Khanchi, J Potgieter, KM Arif, Image-based plant disease identification
chine, in: International Conference on Circuits and Systems in Digital Enterprise by deep learning meta-architectures, Plants (Basel) 9 (11) (2020) 1–23.
Technology, 2018, pp. 1–5. [32] Shree Hari Pushpa, Adarsh Ashok, Diseased leaf segmentation from complex back-
[24] Vidyashree Kanabur, Sunil S. Harakannanavar, Veena I. Puranikmath, Dat- ground using indices based histogram, in: IEEE International Conference on Com-
taprasad Torse, Detection of Leaf Disease Using Hybrid Feature Extraction Tech- munication and Electronics Systems, 2021, pp. 1502–1507.
niques and CNN Classifier, Springer Comput. (2019) 1213–1220. [33] Mohammed Khalid Kaleem, Nitin Purohit, Kassahun Azezew, Smegnew Asemie, A
[25] MH Saleem, J Potgieter, Arif K Mahmood, Plant Disease Detection and Classification modern approach for detection of leaf diseases using image processing and ML based
by Deep Learning, Plants (Basel) 8 (11) (2019) 1–22. SVM classifier, Turkish J. Comput. Math. Educ. 12 (13) (2021) 3340–3347.
[26] Muhammad Amir Nawaz, Tehmina khan, Rana Mudassar Rasool, Maryam Kausar, [34] Thanjai Vadivel, R. Suguna, Automatic recognition of tomato leaf disease using fast
Amir Usman, Tanvir Fatima Naik Bukht, Rizwan Ahmad and Jaleel Ahmad, “Plant enhanced learning with image processing, Taylor Francis, Acta Agricult. Scandinav-
Disease Detection using Internet of Thing (IoT), Int. J. Adv. Comput. Sci. Appl. 11 ica, Sect. B — Soil Plant Sci. 71 (1) (2021) 1–13.
(1) (2020). [35] T.S. M, S.K. S, S.M. S, P.R. Devi, in: Leaf Disease Detection Using Deep Learning”
[27] S. Kumar, K. Prasad, A. Srilekha, T. Suman, B.P. Rao, J.N. Vamshi Krishna, Leaf Dis- IEEE International Conference on Electronics and Sustainable Communication Systems,
ease Detection and Classification based on Machine Learning, in: IEEE International 2021, pp. 1797–1804.
Conference on Smart Technologies in Computing, Electrical and Electronics, 2020, [36] Radhika Bhagwat, Yogesh Dandawate, Comprehensive multilayer convolutional
pp. 361–365. neural network for plant disease detection, Int. J. Adv. Comput. Sci. Appl. 12 (1)
[28] Md.Tariqul Islam, Plant disease detection using CNN model and image processing, (2021) 204–211.
Int. J. Eng. Res. Technol. 09 (10) (2020) 291–297. [37] Rehan Ullah Khan, Khalil Khan, Waleed Albattah, Ali Mustafa Qamar, Image-based
[29] MH Saleem, J Potgieter, KM Arif, Plant disease classification: a comparative evalu- detection of plant diseases: from classical machine learning to deep learning journey,
ation of convolutional neural networks and deep learning optimizers, Plants (Basel) Wireless Commun. Mob. Comput. 2021 (2021) 1–13.
9 (10) (2020) 1–16. [38] K.V Gokula, J. D, R. Pinagadi Venkateswara, V. D, S. K, An automated segmentation
[30] S.B. Gharpankar, Plant Leaf Disease Detection using Convolutional Neural Networks, and classification model for banana leaf disease detection, J. Appl. Biol. Biotechnol.
Int. J. Res. Appl. Sci. Eng. Technol. 8 (11) (2020) 494–499. 2022 (2022) 1–12.

310