Plant Leaf Disease Detection Classification and Diagnosis Using Computer Vision and Artificial Intelligence A Review

Received 4 February 2024, accepted 26 February 2024, date of publication 4 March 2024, date of current version 14 March 2024.
Digital Object Identifier 10.1109/ACCESS.2024.3373001
Plant Leaf Disease Detection, Classification, and

Diagnosis Using Computer Vision and
Artificial Intelligence: A Review
ANUJA BHARGAVA 1 , AASHEESH SHUKLA1 , OM PRAKASH GOSWAMI2 ,
MOHAMMED H. ALSHARIF 3 , PEERAPONG UTHANSAKUL 4 , (Member, IEEE),
AND MONTHIPPA UTHANSAKUL 4 , (Member, IEEE)
1 Department of Electronics and Communication Engineering, GLA University, Mathura 281406, India
2 Department of Electronics, Somaiya Vidyavihar University, Mumbai 400077, India
3 Department of Electrical Engineering, College of Electronics and Information Engineering, Sejong University, Seoul 05006, Republic of Korea
4 School of Telecommunication Engineering, Suranaree University of Technology, Nakhon Ratchasima 30000, Thailand
Corresponding authors: Mohammed H. Alsharif (malsharif@sejong.ac.kr) and Peerapong Uthansakul (uthansakul@sut.ac.th)

This work was supported in part by the Suranaree University of Technology (SUT); in part by Thailand Science Research and Innovation
(TSRI); and in part by the National Science, Research and Innovation Fund (NSRF).
ABSTRACT Agriculture is the ultimate imperative and primary source of origin to furnish domestic income
for multifarious countries. The disease caused in plants due to various pathogens like viruses, fungi, and
bacteria is liable for considerable monetary losses in the agriculture corporation across the world. The
security of crops concerning quality and quantity is crucial to monitor disease in plants. Thus, recognition of
plant disease is essential. The plant disease syndrome is noticeable in distinct parts of plants. Nonetheless,
commonly the infection is detected in distinct leaves of plants. Computer vision, deep learning, few-shot
learning, and soft computing techniques are utilized by various investigators to automatically identify
the disease in plants via leaf images. These techniques also benefit farmers in achieving expeditious and
appropriate actions to avoid a reduction in the quality and quantity of crops. The application of these
techniques in the recognition of disease can avert the disadvantage of origin by a factious selection of disease
features, extraction of features, and boost the speed of technology and efficiency of research. Also, certain
molecular techniques have been established to prevent and mitigate the pathogenic threat. Hence, this review
helps the investigator to automatically detect disease in plants using machine learning, deep learning and few
shot learning and provide certain diagnosis techniques to prevent disease. Moreover, some of the future works
in the classification of disease are also discussed.
INDEX TERMS Deep learning, diagnosis, image processing, machine learning, plant disease.
I. INTRODUCTION than 85% of people confide in agriculture in the world [2].

The agriculture and food organization of the United Nations Accordingly, the food industry and agriculture demand ade-
disclosed that the total number of starved people across the quate farming mechanisms. Further, living creatures accept
world has been growing deliberately since 2015 [1]. The oxygen from plants in the balancing environment by photo-
ongoing estimation reveals that around 680 million people are synthesis. The plant disease that affects the leaves may lead to
starved and they comprise under 9% of the population around death or bad health of the plant and abort to implementation
the world. This indicates an expansion of 10 million in a year of plant food. Also, plant disease has a gigantic effect on
and approximately 120 million in 10 years. Around more developing food crops. In 1845 potato (Irish famine) resulted
in 1.2 million deaths [3]. Some of the laboratory approaches
The associate editor coordinating the review of this manuscript and used for plant disease are immunosorbent enzymes, isother-
approving it for publication was Luca Barletta. mal amplification, and polymerase reaction. Hence, early
2024 The Authors. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License.
VOLUME 12, 2024 For more information, see https://creativecommons.org/licenses/by-nc-nd/4.0/ 37443
A. Bhargava et al.: Plant Leaf Disease Detection, Classification, and Diagnosis
detection, management, and prevention of disease in plants

are precisely essential. Nonetheless, disease detection of
plants in the enormous field is a very complicated task
that involves trained manpower and optical examination of
leaves [4]. Farmers observe the symptoms of diseases on plant
leaves with the naked eye and diagnose plant diseases based
on experience, which is laborious, time-consuming, and
requires special skills [5]. The objective of a disease detection
system is to support non-expert users, i.e. non-pathologist and
non-botanists. Therefore, various automated techniques have FIGURE 1. Phytopathology objectives.
been developed using image processing, machine learning,
deep learning, few short learning are discussed in this review
paper. The agriculture practical applications are not sufficient
using conventional ML approaches due to robustness and
laboratory conditions [6], [7]. The deep learning in recent
times has built exceptional in the classification of images for
plant disease [8], [9]. The approach required a large number
of samples to meet the challenges in time. Also, all the images
are fundamentally annotated by pathologists and ethnolo-
gists. Since these efforts are laborious and time-consuming,
deep learning applications in disease detection are costly [10],
[11]. FSL facilitates learning from machines using a limited
number of labeled databases [12], [13]. The experiment’s
purpose and problem complexity determine the number of
shots required. Various pathogens [14], [15] cause the dis-
ease a diagnosis by DNA (Deoxyribose nucleic acid), PCR
(Polymerase chain reaction), MPG (Modified panchayat mix-
ture), ELISA (Enzyme-linked immunosorbent assay), FISH
(Fluorescence in situ hybridization) and IF (Immunofluo-
rescence) method. Therefore, the objective of this review
paper is to give a comparative analysis of machine learning,
deep learning, few short learnings in plant disease detection FIGURE 2. Subdomains of phytopathology [16].
and also review various segmentation, feature extraction, and
classification techniques. Also, certain molecular diagnosis
tools are reported. III. PLANT DISEASE/TYPES AND SYMPTOMS
The behavior or physiology of the plants in terms of abnor-
mality results in plant disease. It is due to biotic or abiotic
II. PHYTOPATHOLOGY agents as indicated in Fig. 3 [16]. Biotic disease is gen-
The study of plant pathogens, responsible pathogens their dis- erated from infectious agents whereas abiotic disease is
ease, mechanism, control methods of disease and to diminish generated from non-infectious agents. The abiotic disease
their brunt on plants is referred to as Phytopathology [16]. is usually avoidable due to its non-transmissible lesser haz-
As a result, it is a system for dealing with a plant’s life ardous nature. Thus, in this manuscript, biotic disease is
cycle. Phytopathology is a Greek letter word that means examined.
plant (Phyto), disease (Patho), and knowledge (Logo). The • Bacterial Disease- Plant bacterial disease occurs due
basic objectives of phytopathology are a study of the source, to water-soaked results in a little green blemish. These
and causes of plant disease as biotic or abiotic (etiology), lesions augment and then emerge as dead dry blemishes
a mechanism study of development for disease (pathogen- (spots) represented in Fig. 4(a). Example: The foliage
esis), the interaction between plant disease and pathogen consists of water-soaked black blemish or brown leaf
(epidemiology), and development management for reduction spot or halo yellow with equivalent size. The blemish
of radiation and losses control management as shown in appears as dappled beneath dry conditions. Generally,
Fig. 1. Bacterial wilt occurs in brinjal crops due to which the
Phytopathology comprises of subdomain under agriculture whole plant falls [17].
science and consists of elemental learning of microbiology, • Viral Disease- Plant viral disease is the most burden-
physiology, nematology, virology, anatomy, bacteriology, some disease to investigate among all the plant diseases.
mycology, genetic engineering, botany, meteorology, clima- The virus observed on the plant has no indication or
tology, and molecular biology indicated in Fig. 2. it looks similar to herbicide bruise as well as nutrient
37444 VOLUME 12, 2024

FIGURE 3. Classification of plant disease in distinct categories [16].
FIGURE 4. (a) Bacterial blemish [19] (b) Viral Mosaic [20] (c) Late Blight [22] (d) Early Blight [22] (e) Rust.
TABLE 1. Distinct disease in different plants.
failure [17]. The most commonly found viral diseases waterlog shown in Fig. 4(c). This fungus is caused due
are in beetles, leafhoppers, aphids, and whiteflies such as to climate change in the form of wet and dry. As the
mosaic viral disease indicated in green or yellow stripes late blight disease is cultivated, these blemishes get dark
for foliage as shown in Fig. 4 (b). and fungus in white form grows on the surface [18].
• Fungal Disease- Plants fungal disease affects various Alternaria fungus caused by early blight, emerges on
components of plants i.e., sclerotium wilt, common former leaves in the form of tinny brown blemish with
crown rot, stem rust, eyespot (sheath or stem), rust, bull’s eye arrangement in the form of concentric rings as
blight (leaves), ergot (spikes) and carnal bunt, black shown in Fig. 4(d). Rust fungus develops on ripe plant
point (seeds). Phytophthora fungus caused by late blight leaves on the curtailed surface as shown in Figure 4(e).
emerges on former leaves such as gray-green blemish, This blemish becomes black after green-yellow.
VOLUME 12, 2024 37445

TABLE 2. Distinct plants, their disease and responsible pathogen [16].
The above-described and mentioned disease symptoms utilize machine learning (ML), deep learning (DL), few-shot
are distinct and very less compared to other conventional learning (FSL), and soft computing techniques with image
diseases in plants. That means the equivalent disease processing for RGB as well as hyperspectral images. Also,
symptom may arise due to infectious or non-infectious certain molecular techniques have been established to prevent
diseases. Some of the distinct diseases in plants are and mitigate the pathogenic threat.
indicated in Table 1.
Therefore, accurate diagnosis of plant disease is trouble-
A. MACHINE LEARNING USING IMAGE PROCESSING
some for the identification of pathogens using particular
The categorization of disease detection in plants using
symptoms.
machine learning with image processing is done in the
The artificial intelligence techniques, for detecting the per-
following sequential steps: Image Acquisition, Image Pre-
formance of disease detection in plants essentially depend
processing, Image Segmentation, Feature Extraction and
on the feature extraction as well as a classification for dis-
selection, and Classification [32]. In this, firstly images of
ease [27], [28], [29]. These extracted selected features reveal
plants (roots, branches, stems & leaves) are captured. Next,
several disease symptoms. Table 2 represents a manifestation
image pre-processing i.e., noise removal or blur or any
of several typical diseases in plants with specific symptoms.
other distortion is removed using techniques like scaling,
The following Table will help the investigator to identify the
stretching, smoothening, transformation, and rotation. Next,
accurate features which would further lead to high classifica-
segmentation to extract the infected region using techniques
tion accuracy.
like thresholding, and clustering is done. Then this seg-
IV. PLANT DISEASE DETECTION SYSTEM
mented region is used to extract features for classification.
The artificial intelligence techniques are addressed to In this section, the requirement of all these steps is men-
enhance agriculture production by bestowing in plant disease tioned and the approaches proposed in the literature are also
monitoring. Various review research papers have been pub- summarized.
lished [30], [31], some of them focus on particular approaches
while others are centered on particular diseases. A detailed 1) IMAGE ACQUISITION
summarization of plant disease detection, classification, and The initial gait for any machine learning system is image
diagnosis is not available. Therefore, this review focuses acquisition. This incorporates image capture or redeems
on distinct techniques used by distinct researchers that images from repositories. The image quality highly leans on
37446 VOLUME 12, 2024

the camera and its acclimatization which leads to disease TABLE 3. Details of the dataset available used by various researchers.
detection accuracy of the system [33], [34]. The captured
image might dwell in the undesired background, adumbra-
tion, and noise. To improve the accuracy of the system or to
extract the affected part firstly background removal and noise
removal are required. Subsequently, other than RGB images
from normal cameras, certain specific cameras are used
to capture hyperspectral, thermal, and fluorescent images.
Various researchers use distinct datasets for plant disease
detection systems as indicated in Table 3. The presence of
shadow, light, insufficient lighting, and intricated background
is burdensome due to uncontrolled/real-time images as com-
pared to laboratory surroundings conditions. The quality of
the image captured robustly depends upon the techniques and
equipment used. Consequently, the system performance is
influenced by the image acquisition stage.
2) IMAGE PRE-PROCESSING
The crucial step in a machine learning system is image pro-
cessing. It helps to improve the quality of the image that
is degraded due to distortions, turbulence, or shadow [56].
Majorly, the dataset is captured in uncontrolled environments
or real-time situations. Therefore, the preprocessing step is
necessary before feature extraction enlarges the estimated
accuracy of the system which results in reduced process-
ing time. The process also reduces the processing time by
applying operations such as crop, and resize. The literature
utilizes image enhancement, cropping, and elimination of
background as extensively used preprocessing techniques.
These operations applicability depends on the image quality.
Various researchers use unique techniques for preprocessing
are indicated in Table 4. Image preprocessing along with
augmentation is used to create more images of the data on the
existing data for several deep learning techniques that require
large amounts of image data. The processing augmentation
techniques are noise injection, flipping of image, gamma
correction, scaling, rotation, resize, cropping, random shift,
zoom, and image transformation such as contrast enhance-
ment and brightness.
3) IMAGE SEGMENTATION
The region of interest or undesired region can be segre-
gated using image segmentation. This approach intends to
separate the part having a deformity. This analysis of the
representation of the image is simple and further meaningful
to discriminate the affected and unaffected parts. It is the most
important technique among machine learning techniques as
critical issues and techniques in the analysis of an image.
The several challenges faced such as erroneous boundaries
of the affected region, shadow, illumination, and complex and neural network. Generally, computational techniques per-
background [74]. The segmentation process is classified into form better than conventional techniques. The segmentation
two main divisions: first is conventional methods such as process is crucial for the extraction of features. Various
thresholding, region growing and edge detection other is com- researchers use unique techniques for segmentation are indi-
putational techniques such as fuzzy logic, genetic algorithm, cated in Table 5.
VOLUME 12, 2024 37447

TABLE 4. Details of pre-processing technique used by various TABLE 5. Details of segmentation technique used by various researchers.
researchers.
before classification. The features are advantageous to ana-

lyze objects and to allow the class label to an object.
Generally, in the literature color, shape, and texture features
are utilized for recognition. The classification of plant disease
detection systems is enormously dependent upon the tech-
niques used for image characteristic extraction [90]. From the
literature, the spotted area segmentation and texture, color &
shape feature extraction are of huge importance. Several plant
disease detection researchers have used distinct convenient
feature extraction techniques depending on texture, color &
shape. Therefore, determining the correct features is a major
issue in the detection of disease due to similarity in disease
characteristics.
In computer vision/ machine learning applications, the
selection of relevant features is also important to prevent over-
4) FEATURE EXTRACTION fitting and huge computational costs. Therefore, in computer
To discriminate one part of an image from another part, vision and machine learning applications feature selection
feature attribute extraction is one of the most important approaches are selected to find better relevant features from
techniques in computer vision/ machine learning systems the extracted features. Some of the techniques such as PCA,
37448 VOLUME 12, 2024

genetic algorithm, and particle swarm optimization are used The present scenario of literature during the last 12 years
for feature selection. The literature utilizes various feature using different feature extraction techniques for distinct crops
extraction techniques indicated in Table 6. By examining is represented in Fig 5. There are still various techniques to
these approaches, an attempt is made to find the better-suited consider for plant disease detection using machine learning
technique. and computer vision techniques. The crop selection for such
research is utilized by the availability of experts and data sets.
TABLE 6. Details of feature extraction technique used by various
researchers. 5) CLASSIFICATION
The utmost extrusive stage in computer vision/ machine
learning is classification. The attainment of this step builds
upon antecedent steps such as acquisition, preprocessing,
and feature extraction/ selection. The manuscript is focused
on disease detection in plants, the above steps are used for
detection based on category and symptoms. In this system,
a dataset is trained first and then used to classify the test
image as healthy/defective. ‘‘Machine learning (ML) is an
application of artificial intelligence (AI) which provides you
a system an ability to automatically learn, improve from
experiences without being explicitly programmed and which
can make decisions [124].’’ ML techniques have been broadly
distinguished as Supervised ML and Unsupervised ML. Dur-
ing the training process the labeled and unlabeled data work
for supervised and unsupervised techniques respectively. The
semi-supervised ML is another special supervised learning
that works on both labeled and unlabeled data during the
training process. In the literature, various classification tech-
niques are utilized as indicated in Fig. 6.
Generally, support vector machine (SVM), artificial neural
network (ANN), and k-nearest neighbor (k-NN) have been
adapted for machine learning techniques [32]. Also, in a
few studies vegetation index and fuzzy logic feature-based
techniques are utilized. Distinct models of ANN like feed-
forward NN, error backpropagation, multilayer perceptron,
probabilistic neural network, and self-organizing map have
been widely utilized for plant disease detection. Table 7
indicates the quality investigation of plant disease detection
based on distinct classification techniques used by distinct
researchers.
B. DEEP LEARNING USING IMAGE PROCESSING

Deep learning (DL) was first proposed in 1943 [146] as a hier-
archy of machine learning for the detection of objects [147],
[148], [149], [150], classification of images as well as pro-
cessing of natural language [151], [152], [153]. DL forms
high-level features by combining low-level features for dis-
tributed features discovering and sample data attributes. The
ability of generalization and accuracy are enhanced com-
pared to traditional ML techniques for target detection and
image recognition. The development of DL is in three phases.
Phase 1, during 1943-1969 was a linear model of neural
network for classification. Phase 2, during 1986-1988 was
a non-linear mapping of multi-layer perceptron for the BP
(Backpropagation) algorithm. Phase 3, from 2006 to the
present, has a ReLu activation function, ImageNet recogni-
tion, and deep learning model-AlexNet. The DL methods are
VOLUME 12, 2024 37449

FIGURE 5. Different feature extraction techniques for distinct crops during last 12 years [16].
used for selecting features automatically. The techniques pro- interface programming application such as deployment of
pose high-level features from the combination of low-level the network, model support, and duplication of code is pro-
features. Subsequently, multifold advanced DL techniques vided by the DL framework [157]. Currently, the widely used
proposed are shown in Fig. 7. Presently, the preeminent frameworks are Café, PyTorch, and Keras. The elemental
neural networks are convolutional neural networks, multi- steps of DL are indicated in Fig. 8 where CNN (Convolu-
layer perceptron, and recurrent neural networks. The CNN tional Neural Network) is used to incorporate features via
establishes a classification model that incorporates AlexNet, convolutional, max-pooling, and fully connected layers. The
GoogleNet, VGGNet, MobileNet, ResNet, and Efficient- features are extracted via the convolutional layer whereas
Net. AlexNet is the early model to adopt GPU devices texture and edges are extracted via shallow convolutional
for network training. The local normalization and rectified layer. The semantic information and complex texture features
linear units are used as activation functions. Then VGG are extracted using the middle layer. The high-level semantic
& GoogleNet use a few kernels with low memory [154]. features are extracted using a deep layer. The max pool
In 2015, ResNet proposed by Microsoft uses connections and layer is used to hold the critical information in an image.
balance blocks [155]. In 2017, MobileNet was proposed by Lastly, the fully connected layer is used to classify high-
Google teams for embedded and mobile applications [156]. level features. Therefore, the implementation of DL requires
In 2019, EfficientNet was introduced by Google teams that the fundamental steps which are discussed in detail in the
use efficient coefficients to scale width, depth, and resolution. subsequent section.
The deep network may not be used in plant disease detection
since the basic models like VGG16 and AlexNet can fit the 1) IMAGE ACQUISITION
substantial accuracy needed. The C/C++, Python program- The data acquired must be accurate to obtain a perfect model.
ming language can be used in the DL model. The various The DL database consists of training, validation, and test
37450 VOLUME 12, 2024

FIGURE 6. Distinct classifiers are analyzed in research for the detection of disease in plants.
FIGURE 7. Hierarchy of DL techniques.
sets. The model is learned by the training set, hyperparam- small datasets in deep learning. Therefore, the expansion
eters are adjusted during validation and the performance of the dataset is essential for plant leaf disease detec-
evaluation is done by the test set [147], [158]. Various tion with deep learning. The practical application and the
publicly available datasets are indicated in Table 8. These basic requirement can be removed using data augmentation.
publicly available dataset is from two websites BIFROST For example, the fundamental manifestation of plant dis-
(https://datasets.bifrost.ai/, accessed on 15 November 2023) ease is color. The original image color will not change by
and Kaggle (https://www.kaggle.com/datasets, accessed image enhancement. The two ways for data augmentation
on 12 November 2023) which can be used for classification. are conventional augmentation and generating an adversarial
In the literature of DL technique applied to plant disease clas- network.
sification, the most used public datasets are PlantVillage [33], Conventional techniques can boost the dataset via rotation,
[159] and Kaggle [160], many others also collect their own saturation adjustment, or symmetry of the mirror. Some of
datasets [161], [162]. the advisory use augmentation techniques are AugMix [163],
Fast AutoAugment [162], CutMix [164], population-based
2) IMAGE AUGMENTATION augmentation [165] and Rand Augment [166]. The flaws of
The plant leaf disease labeling, collection, and detection tech- this technique are deficient diversity, unevenness, and meager
nique requires a huge number of datasets with financial and quality. The data augmentation technique is used by various
manpower resources. It is burdensome to collect short-onset researchers to improve the efficiency of the system as men-
period plants. The recognition rate will be destitute for tioned in Table 9.
VOLUME 12, 2024 37451

TABLE 7. Details of classification technique used by various researchers.
37452 VOLUME 12, 2024

TABLE 7. (Continued.) Details of classification technique used by various researchers.
VOLUME 12, 2024 37453

TABLE 7. (Continued.) Details of classification technique used by various researchers.
FIGURE 8. Elemental steps of deep learning.
TABLE 8. Bulky publicly available dataset used in deep learning. TABLE 9. Various literature regarding data augmentation.
Goodfellow et al. 2014 [167] propose a generating model

known as Generate Adversarial Network (GAN). Conse-
quently, huge variations of GAN have developed sepa- This model primarily contains a generator and discriminator
rately like CGAN, DCGAN, LAPGAN, PGGAN, InfoGAN, as shown in Fig. 9. The GAN technique is used by various
F-GAN, WGAN, LeakGAN, SeqGAN. The objective is to researchers to improve the efficiency of the system as men-
generate samples of a dataset with the equivalent attributes. tioned in Table 10.
37454 VOLUME 12, 2024

TABLE 11. Details of DL technique used by various researchers.
FIGURE 9. Generate adversarial network architecture.
TABLE 10. Various literature regarding data augmentation.
3) IMAGE CLASSIFICATION
Recently, several fruitful applications of DL for the classifica-
tion of disease in plants have been established. Nonetheless,
transparency and interpretability are curtailments in classi-
fication using DL. The mechanism consists of a black box
beyond one minutiae or explanation. The disease detection
system must reveal the disease, and symptoms as well as
achieve high accuracy. Various researchers utilize distinct
techniques using DL as indicated in Table 11.
C. HYPERSPECTRAL IMAGING USING IMAGE

PROCESSING
The disease in plants is burdensome sometimes via computer
vision systems or naked eyes due to pathogens’ growing
process [221]. The hyperspectral imaging sensor is primarily
fixed in the electromagnetic spectrum range of infrared and
visible (400 nm ∼ 2500 nm). These sensors attain shadowy
information from a wide range of bands [222]. This range
is eminently sensitive to leaf variation caused by distinct leaf
disease such that antecedent pathology detection can be done.
Consequently, hyperspectral imaging targeted plant disease
detection at the early stage. Wang et al. 2019 [149] propose
a GAN for hyperspectral imaging as OR-AC-GAN (Outlier
Removal—Auxiliary Classifier-GAN) for tomato leaf dis-
ease detection and achieve 96.25% accuracy as shown in
VOLUME 12, 2024 37455

TABLE 11. (Continued.) Details of DL technique used by various TABLE 12. Details of hyperspectral image classification used by various
researchers. researchers.
An extensive number of modern developments in different

pathogen systems using distinct group of extremely sensitive
sensors and various data analysis pipeline have been sum-
marized in Fig 11 and Table 13. It overviews the current
sensors technologies used for automatic identification and
detection of plants. These sensors can be implemented in
agriculture applications of plants. Depending on the scale,
distinct platforms of plants can be managed and observed.
RGB Imaging: RGB (Red Green Blue) is the simple source
og digital images for disease detection and identification. The
technical parameters such as photo sensor for light sensitiv-
ity, optical focus and spatial resolution has been improved
significantly [238].
Multi and Hyperspectral Sensors: Multispectral sensors
assess the spectral imformation of object in various wave-
bands [239]. It also provide RGB and near infrared wave-
bands [240]. Hyperspectral sensors provides spatial and
spectral information from the object. The spatial resolution
depends upon the object and the sensor [241].
Thermal Sensors: Infrared thermography (IRT) assess
plant temperature and is correlated with the status of plant
water, the microclimate in crop stand [242]. IRT can be used
in plant science at distinct spatial and temporal scales from
small scale applications [238].
Fig. 10. Various researchers use distinct techniques for HSI Fluorescence Sensors: Chlorophyll fluorescence sensors
plant disease detection are illustrated in Table 12. are active sensors with laser light or LED source that
37456 VOLUME 12, 2024

assess photo synethitic transfer [243]. It combine with image includes data, models, and optimization to tune the samples.
analysis technique have been useful for quantification and The training phase leads to creating source tasks activated by
discrimination of fungal diseases. obtaining multiple tasks from single or multiple datasets as
represented in Fig. 12. The machine learning model is com-
TABLE 13. Details of OPTICAL SENSORS used by various researchers. bined with these source tasks to train the system. The most
commonly available dataset for plant recognition systems
consists of several thousands of images. The performance
of the system depends on the quantity as well as the quality
of images. Therefore, FSL is the best possible feasible solu-
tion for a confined number of samples [253], [254]. Various
researchers use distinct techniques for small samples-based
plant leaf disease detection are illustrated in Table 14.
Various applications for FSL to abate parameters are men-
tioned below.
i. Embedding- It features an extraction technique also
known as data transformation benefitted for low dimen-
sion reduction (higher dimensional map to lower
dimension). This includes pre-trained CNN with Ima-
geNet for extraction of features (reduce training time)
and SVM for classification. SVM uses distinct kernels
to handle complex systems [268].
ii. Multitask Learning- The process includes a shared
model to train multiple tasks generally classified as
hard parameter sharing and soft parameter sharing.
TABLE 14. Details of FEW SHOT image classification used by various
researchers. In HPS certain parameters are shared to the entire task
and other parameters learn precise features. In SPS
each parameter is shared with individual tasks. This
multitask learning incorporates all the models that
can be trained for all the tasks rather than a single
task [269].
iii. Transfer Learning- The process was initiated to
accomplish a diminished training period by reusing
gathered knowledge [270].
iv. Meta-Learning- This refers to optimization and per-
formance improvement via multiple-task experience.
It is also known as support sets that learn from mul-
tiple sources instead of a single source (conventional
machine) alongside FS to adequately develop a model
for clarifying query sets.
The meta-learning and multitasking learning are generally
unstable and marginally upgraded by a few proportions.
Transfer learning is the most superior and stable to train
the model from scratch. These approaches are computa-
tionally costly compared to the degree of performance for
enhancement. The future work led to enhancing existing
FSL algorithms to cooperate with plant disease character-
istics recognition. Further comparative and comprehensive
studies are done for the performance evaluation of each FSL
D. FEW SHOT LEARNING (FSL) USING IMAGE approach.
PROCESSING
A new model of machine learning that facilitates learning E. MOLECULAR DIAGNOSIS TECHNIQUES
from machines using a limited number of labeled databases is Pest and disease may heighten the total food production
known as FSL [252]. The experiment’s purpose and problem losses by 35%. Globally, under quarantined pathogens,
complexity determine the number of shots required. FSL some of the austere plant diseases are nematodes, parasitic,
uses datasets with prior knowledge for training. This process oomycetes, viruses, bacteria, and fungi.
VOLUME 12, 2024 37457

FIGURE 10. Segmentation comparison (a) Direct CNN (b) AC-GAN (c) OR-AC-GAN.
FIGURE 11. Overview of sensor technologies [238].
FIGURE 12. Approach to imitate multiple sources.
Fang and Ramasamy 2015 [15] reported early disease diag- and detects the particular substance in the sample, and
nosis is critical to overcome outbreaks of disease. Scientists provokes a color change that expands antibodies with
utilize laboratories and certain advanced facilities compared enzymes [271]. ELISA is not very sensitive compared
with traditional techniques resulting in reliable diagnosis to other methods [272] but is labor-intensive and sen-
methods. Nonetheless, these approaches performed better sitive [273]. Despite antibodies like polyclonal being
only with severe infection in plants. Therefore, to bridge adequate to distinguish pathogens, nevertheless, these
this gap certain adequate molecular diagnosis techniques like are not ample specific. One other antibiotic monoclonal
management control on decisions or tools for plant disease is innumerable specific but further costly. As ELISA
overcome the traditional techniques. The emerging molecular is tremendously sensitive to viruses results in the suc-
technologies to diagnose plants due to the necessity of man- cessful development of antibodies [274]. The technique
aging diseases are as follows. will be inadequate to detect bacteria due to low sen-
i. Enzyme-linked immunosorbent assay (ELISA): sitivity [15]. Therefore, pathogens bacteria, and fungi
This approach is widely employed as a diagnosis for rice disease are not effective using ELISA. Also
device in plant and chemical pathology. It identifies due to inadequacy of specificity, ELISA is not able
37458 VOLUME 12, 2024

FIGURE 13. Illustrative real-time PCR schematic [286].
FIGURE 14. Amplification of LAMP [283].
FIGURE 15. Biosensors & its process schematic illustration [286].
to discriminate several strains with luminous specific (PCR) established in the mid-80s [276]. The tech-
symptoms [275]. nique requires various components and reagents which
ii. Conventional PCR: The most valuable achievement in incorporate monovalent cation, primers, buffer solu-
molecular biology is the Polymerase Chain Reaction tion, triphosphate deoxynucleotide, and polymers.
VOLUME 12, 2024 37459

The technique anticipates electrophoresis used for the device includes cost effectiveness, sensitivity, and
the detection of plant pathogens. The nucleic acid simplicity.
evocation technique affects the PCR sensitivity and vi. Next Generation Sequencing (NGS): Generally, the
(Dioxyribo Nucleic Acid) DNA target variability [277]. sequencing mechanism depends on molecular detec-
The highly sensitive nature of PCR results in DNA tion approaches [288]. The first generation of commer-
amplification. The PCR limits contamination suscep- cial and laboratory sequencing was employed due to
tibility and inadequacy of robustness for a long time poor radioactivity and improved efficiency [289]. The
with false positive results. approach is ineffective and tedious for large sequences
iii. Real-Time PCR: A significantly revolutionary tech- and efficient for short sequences. The NGS is also
nique developed by Mullis Karry in 80 is also known known as a parallel massive sequence capture sequence
as a quantitative polymerase chain reaction. This tech- in less time, low cost, and high throughput [290]. The
nique diminishes the cross-contamination risk with unique biology molecular opportunities are incorpo-
most precipitate-sensitive detection [278]. The infected rated via NGS such as mela genomics [291], single-cell
tissues and vectors can easily detect DNA pathogens sequencing, sew protein [292], and whole genome
using RT-PCR [279]. This technique points exclusively sequencing [293]. The approach limits to diagnosis of
to known genes [280] with sequence data indicated in helm pathogen [294], [295]. Advanced NGS adopts
Fig. 13. colony sequencing, parallel massive sequencing, and
iv. Lamp: The most common alternative to RTPCR is a pyrosequencing by legation nucleotide [287].
lamp due to its cost-effectiveness, simplicity, rapid-
ity, and practicality. Under isothermal elaboration, V. LIMITATIONS
lamps recognize HBV (Hepatitis B Virus) using certain While plant disease detection has been analyzed automati-
specific sequences of elaborated DNA [281]. Lamp cally, there are several limitations in each step or process.
contains four distinct primers two for inner and outer The samples are very difficult to obtain for particular dis-
each (Forward inner primer, backward inner primer, eases [37], [38], [39]. It is observed that system performance
forward outer primer, backward outer primer) as indi- is immensely influenced by the type of dataset used i.e.
cated in Fig. 14. A loop is formed for high displacement laboratory controlled or real-time dataset. The data obtained
and auto strand DNA for 60 minutes at 65 degrees C in an uncontrolled environment results in increased sys-
isothermal situation [282]. As indicated in the figure tem complexity but it is of huge significance in agriculture
the process of corporate initial material production, development and more adequate in today’s research. Sev-
amplification, recycling, and elongation [283]. Lamp eral difficulties have been observed during the extraction of
detects amplicon using the detection of colorimet- features [106]; similarity in the infected area leads first to
ric, fluorimeter, and turbidimeter [284]. The technique the extraction of inappropriate features and then false clas-
incorporates the formation of primer to cultivate the sification based on irrelevant feature matching. Thus, it is
65-degree C temperature [285]. The technique is highly necessary to select a set of features because each feature has a
sensitive for elaboration therefore proper practice and different importance level. Various classification techniques
handling are enforced to escape the contamination risk. used for plant disease detection are SVM, ANN, Naïve Bayes,
v. Biosensor: An ample range of challenges and backpropagation neural network, decision tree, and k-nearest
problems in pharmacology medicine, security, food neighbor [123], [124], [125], [126]. Despite these classifiers
safety, and agriculture can be resolved using convolutional neural networks have been explored for huge
biosensors. Clark and Lyons [287] develop biosen- samples (thousands and more) more accurately. However, the
sors via measurement of glucose using hydro- overfitting of CNN is seen as a major issue in systems about
gen peroxide or oxygen electrochemical detection. plant disease detection using deep neural networks [195].
Vigneshvar et al. 2016 further evolved biosensors Various approaches have been observed which can improve
into innovative approaches like nanotechnology, and the efficiency of convolutional neural network-based classi-
electrochemistry. fication systems for multiple cultures.
‘‘According to International Union of Pure and Applied According to the literature, the proposed system has a set
Chemistry recommendations (1999), a biosensor is of specifications to be fulfilled necessarily, if any specifi-
described as a device that deploys the integration of cation is unconsidered then the system proposed provides
specific biochemical reactions mediated by isolated inaccurate disease detection. Therefore, a versatile system
enzymes, immune systems, tissues, organelles or whole must be designed by a researcher with an adjustable set of
cells to detect chemical compounds, commonly by specifications instead of a constant. The actual use of machine
electrical, thermal or optical signals’’. These analytical learning has been disturbed due to overfitting. Hence a highly
appliances are constituted of the transducer coupled generalized and adaptable system is to be developed for plant
with a biorecognition component and disciple into the disease detection in various cultures. Machine learning is the
analytical signal as shown in Fig. 15. The advantage of most popular domain due to the availability of techniques
37460 VOLUME 12, 2024

and tools, but these resources must not be compromised for an automated plant disease recognition and classification is
accuracy. reviewed using machine learning, deep learning, and few-shot
learning. The review presents various well-known approaches
VI. CHALLENGES such as acquisition, preprocessing, segmentation, feature
From the literature review, it can be found that one of the chal- extraction, and classification. Despite RGB images some
lenges facing plant disease detection is the lack of experts to of the studies use hyperspectral imaging for plant leaves.
annotate accurately. The issue arises when experts cannot dif- The hyperspectral does not require a labeled database and
ferentiate correctly between dead plants and disease-infected microscopic symptoms are easily detected. These automated
plants. This process requires experienced and expert profes- techniques contribute a manifestation of research in accessi-
sionals to detect diseases in plants that are costly and difficult, ble time. Also, the diagnosis molecular tools with the latest
especially for rare or new diseases. Furthermore, using DL techniques for disease detection are reviewed. The entire
techniques for modeling, tuning, and resource training is techniques indicated in this paper are gratuitous to the sensi-
another important challenge. The shallow architectures are tive, specific, and rapid detection of plants. Moreover, in the
best suited for small datasets. The most recent models for future to detect disease in plants, the combination of severe
disease detection offer another angle to consider in building a side program and client mobile as well as electrophysiology
model for plant disease detection. The ML, DL, and FSL are would be exceptional future research guidance.
to be considered to enhance the detection models. This paper
recommends that future research on plant disease detection, REFERENCES
classification, and quantification of disease will improve mart [1] The State of Food Security and Nutrition in the World 2023, FAO; IFAD;
agriculture. UNICEF; WFP; WHO, Geneva, Switzerland, Jul. 2023.
Various issues and factors that may affect plant disease [2] T. D. March. State of Agriculture in India. Accessed: Aug. 9, 2023.
[Online]. Available: https://prsindia.org/files/policy/policy_analytical_
classification and identification are listed below: reports/State%20of%20Agriculture%20in%20India.pdf
(i) The plant disease detection system performance mainly [3] D. P. Hughes and M. Salathe, ‘‘An open access repository of images on
depends on the feature extraction and classification technique plant health to enable the development of mobile disease diagnostics,’’
2015, arXiv:1511.08060.
used [144]. [4] S. Verma, A. Chug, and A. P. Singh, ‘‘Prediction models for identification
(ii) Most of the literature reviewed utilizes Plant Village and diagnosis of tomato plant diseases,’’ in Proc. Int. Conf. Adv. Comput.,
Dataset i.e. laboratory images rather than real time images. Commun. Informat. (ICACCI), Sep. 2018, pp. 1557–1563.
[5] S. Sankaran, A. Mishra, R. Ehsani, and C. Davis, ‘‘A review of advanced
The classifier performance heavily depends on the dataset techniques for detecting plant diseases,’’ Comput. Electron. Agricult.,
used for testing and training [144]. vol. 72, no. 1, pp. 1–13, Jun. 2010.
(iii) The real field images may have complex backgrounds; [6] C. Janiesch, P. Zschech, and K. Heinrich, ‘‘Machine learning and deep
therefore, segmentation of affected areas is difficult which learning,’’ Electron. Markets, vol. 31, no. 3, pp. 685–695, Apr. 2021, doi:
10.1007/s12525-021-00475-2.
affects the performance of the system [296]. [7] K. Kc, Z. Yin, D. Li, and Z. Wu, ‘‘Impacts of background removal on
(iv) The plant has a deficiency in nutrients [74] and con- convolutional neural networks for plant disease classification in-situ,’’
tamination at an early stage of development [8]. Agriculture, vol. 11, no. 9, p. 827, Aug. 2021, doi: 10.3390/agricul-
ture11090827.
(v) The infected area estimation and severity man- [8] J. Liu and X. Wang, ‘‘Plant diseases and pests detection based on
agement with disease detection may be used to control deep learning: A review,’’ Plant Methods, vol. 17, no. 1, pp. 1–18,
pesticides [188]. Dec. 2021.
[9] L. C. Ngugi, M. Abelwahab, and M. Abo-Zahhad, ‘‘Recent advances
(vi) The real-time illness efficiency on constrained devices in image processing techniques for automated leaf pest and disease
could be considered [189]. recognition—A review,’’ Inf. Process. Agricult., vol. 8, no. 1, pp. 27–51,
(vii) Various diseases may create distinct manifestations Mar. 2021.
[10] P. Chinmayi, L. Agilandeeswari, and P. Manoharan, ‘‘Survey of image
simultaneously. Therefore, it will be difficult to identify and processing techniques in medical image analysis: Challenges and
combine hybrid symptoms. methodologies,’’ in Proc. Int. Conf. Soft Comput. Pattern Recognit.,
(viii) The hyperparameters tuning and selection have the Aug. 2017, pp. 460–471, doi: 10.1007/978-3-319-60618-7_45.
potential to have a powerful influence on performance [196]. [11] I. A. Adeyanju, O. O. Bello, and M. A. Adegboye, ‘‘Machine learning
methods for sign language recognition: A critical review and anal-
(ix) Due to the characteristic of disease uniformity and ysis,’’ Intell. Syst. Appl., vol. 12, Nov. 2021, Art. no. 200056, doi:
attribute selection, disease identification systems face signif- 10.1016/j.iswa.2021.200056.
icant challenges [297]. [12] M. Fink, ‘‘Object classification from a single example utilizing class
relevance metrics,’’ in Proc. 17th Int. Conf. Neural Inf. Process. Syst.,
2004, pp. 449–456.
VII. CONCLUSION [13] L. Fei-Fei, R. Fergus, and P. Perona, ‘‘One-shot learning of object
The manifestation of imminent pathogens plants extends to categories,’’ IEEE Trans. Pattern Anal. Mach. Intell., vol. 28, no. 4,
pp. 594–611, Apr. 2006.
become a considerable threat to food safety, the ecosys-
[14] F. Martinelli, R. Scalenghe, S. Davino, S. Panno, G. Scuderi, P. Ruisi,
tem, and the global economy. Moreover, crucial factors P. Villa, D. Stroppiana, M. Boschetti, L. R. Goulart, and C. E. Davis,
like mobility globalization, vectors, changes in climate, and ‘‘Advanced methods of plant disease detection. A review,’’ Agron. Sus-
tain. Dev., vol. 35, no. 1, pp. 1–25, 2015.
evolution in pathogens emboldened the advancement of inva-
[15] Y. Fang and R. Ramasamy, ‘‘Current and prospective methods for
sive pathogen plants. An adequate automated approach is plant disease detection,’’ Biosensors, vol. 5, no. 3, pp. 537–561,
eminently desired to bury losses in agriculture. Therefore, Aug. 2015.
VOLUME 12, 2024 37461

[16] V. K. Vishnoi, K. Kumar, and B. Kumar, ‘‘Plant disease detection using [37] K. Karadag, M. E. Tenekeci, R. Tasaltin, and A. Bilgili, ‘‘Detection
computational intelligence and image processing,’’ J. Plant Diseases of pepper fusarium disease using machine learning algorithms based
Protection, vol. 128, no. 1, pp. 19–53, Aug. 2020, doi: 10.1007/s41348- on spectral reflectance,’’ Sustain. Comput., Informat. Syst., vol. 28,
020-00368-0. Dec. 2020, Art. no. 100299.
[17] Evaluations of Brinjal Germplasm for Resistance to Fusarium Wilt [38] S. Coulibaly, B. Kamsu-Foguem, D. Kamissoko, and D. Traore, ‘‘Deep
Disease. Accessed: Aug. 9, 2023. [Online]. Available: https://www. neural networks with transfer learning in millet crop images,’’ Comput.
ijsrp.org/research-paper-0717.php?rp=P676604 Ind., vol. 108, pp. 115–120, Jun. 2019.
[18] P. Adhikari, Y. Oh, and D. Panthee, ‘‘Current status of early blight [39] X. E. Pantazi, D. Moshou, and A. A. Tamouridou, ‘‘Automated leaf
resistance in tomato: An update,’’ Int. J. Mol. Sci., vol. 18, no. 10, p. 2019, disease detection in different crop species through image features anal-
Sep. 2017. ysis and one class classifiers,’’ Comput. Electron. Agricult., vol. 156,
[19] S. Zhang, S. Zhang, C. Zhang, X. Wang, and Y. Shi, ‘‘Cucumber leaf pp. 96–104, Jan. 2019.
disease identification with global pooling dilated convolutional neural [40] A. Fuentes, S. Yoon, S. Kim, and D. Park, ‘‘A robust deep-learning-
network,’’ Comput. Electron. Agricult., vol. 162, pp. 422–430, Jul. 2019. based detector for real-time tomato plant diseases and pests recognition,’’
Sensors, vol. 17, no. 9, p. 2022, Sep. 2017.
[20] J. Kianat, M. A. Khan, M. Sharif, T. Akram, A. Rehman, and T. Saba,
[41] S. Shrivastava and D. S. Hooda, ‘‘Automatic Brown spot and frog eye
‘‘A joint framework of feature reduction and robust feature selection
detection from the image captured in the field,’’ Amer. J. Intell. Syst.,
for cucumber leaf diseases recognition,’’ Optik, vol. 240, Aug. 2021,
vol. 4, no. 4, pp. 131–134, 2014.
Art. no. 166566.
[42] J. Parraga-Alava, K. Cusme, A. Loor, and E. Santander, ‘‘RoCoLe:
[21] M. Agarwal, S. Gupta, and K. K. Biswas, ‘‘A new Conv2D model with
A robusta coffee leaf images dataset for evaluation of machine learn-
modified ReLU activation function for identification of disease type and
ing based methods in plant diseases recognition,’’ Data Brief, vol. 25,
severity in cucumber plant,’’ Sustain. Computing: Informat. Syst., vol. 30,
Aug. 2019, Art. no. 104414, doi: 10.1016/j.dib.2019.104414.
Jun. 2021, Art. no. 100473.
[43] K. Tian, J. Li, J. Zeng, A. Evans, and L. Zhang, ‘‘Segmentation of tomato
[22] V. K. Shrivastava, M. K. Pradhan, S. Minz, and M. P. Thakur, ‘‘Rice plant leaf images based on adaptive clustering number of K-means algorithm,’’
disease classification using transfer learning of deep convolution neural Comput. Electron. Agricult., vol. 165, Oct. 2019, Art. no. 104962, doi:
network,’’ Int. Arch. Photogramm., Remote Sens. Spatial Inf. Sci., vol. 42, 10.1016/j.compag.2019.104962.
pp. 631–635, Jul. 2019.
[44] M. Zhang and Q. Meng, ‘‘Automatic citrus canker detection from leaf
[23] J. Chen, D. Zhang, A. Zeb, and Y. A. Nanehkaran, ‘‘Identification of rice images captured in field,’’ Pattern Recognit. Lett., vol. 32, no. 15,
plant diseases using lightweight attention networks,’’ Exp. Syst. Appl., pp. 2036–2046, 2011.
vol. 169, May 2021, Art. no. 114514. [45] R. Pydipati, T. F. Burks, and W. S. Lee, ‘‘Identification of citrus disease
[24] H. Sun, L. Zhai, F. Teng, Z. Li, and Z. Zhang, ‘‘qRgls1. 06, a major QTL using color texture features and discriminant analysis,’’ Comput. Electron.
conferring resistance to gray leaf spot disease in maize,’’ Crop. J., vol. 9, Agricult., vol. 52, nos. 1–2, pp. 49–59, Jun. 2006.
pp. 342–350, Apr. 2021. [46] A. N. I. Masazhar and M. M. Kamal, ‘‘Digital image processing technique
[25] K. P. Ferentinos, ‘‘Deep learning models for plant disease detection for palm oil leaf disease detection using multiclass SVM classifier,’’
and diagnosis,’’ Comput. Electron. Agricult., vol. 145, pp. 311–318, in Proc. IEEE 4th Int. Conf. Smart Instrum., Meas. Appl. (ICSIMA),
Feb. 2018. Nov. 2017, pp. 1–6.
[26] A. Abbas, S. Jain, M. Gour, and S. Vankudothu, ‘‘Tomato plant disease [47] A. S. Deshapande, S. G. Giraddi, K. G. Karibasappa, and S. D. Desai,
detection using transfer learning with C-GAN synthetic images,’’ Com- ‘‘Fungal disease detection in maize leaves using Haar wavelet features,’’
put. Electron. Agricult., vol. 187, Aug. 2021, Art. no. 106279. in Information and Communication Technology for Intelligent Systems.
[27] H. R. Kappali, K. M. Sadyojatha, and S. K. Prashanthi, ‘‘Com- Singapore: Springer, 2019, pp. 275–286.
puter vision and machine learning in paddy diseases identification and [48] J. D. Pujari, R. Yakkundimath, and A. S. Byadgi, ‘‘SVM and ANN based
classification: A review,’’ Indian J. Agricult. Res., vol. 10, pp. 1–5, classification of plant diseases using feature reduction technique,’’ Int. J.
Mar. 2023. [Online]. Available: https://www.arccjournals.com/journal/ Interact. Multimedia Artif. Intell., vol. 3, no. 7, p. 6, 2016.
indian-journal-of-agricultural-search/A-6061 [49] S. Abed and A. A. Esmaeel, ‘‘A novel approach to classify and detect
[28] D. S. Joseph, P. M. Pawar, and R. Pramanik, ‘‘Intelligent plant disease bean diseases based on image processing,’’ in Proc. IEEE Symp. Comput.
diagnosis using convolutional neural network: A review,’’ Multime- Appl. Ind. Electron. (ISCAIE), Apr. 2018, pp. 297–302.
dia Tools Appl., vol. 82, no. 14, pp. 21415–21481, Oct. 2022, doi: [50] M. Azadbakht, D. Ashourloo, H. Aghighi, S. Radiom, and
10.1007/s11042-022-14004-6. A. Alimohammadi, ‘‘Wheat leaf rust detection at canopy scale under
[29] F. Ahmed, M. Taimur Ahad, and Y. Rayhan Emon, ‘‘Machine learning- different LAI levels using machine learning techniques,’’ Comput.
based tea leaf disease detection: A comprehensive review,’’ 2023, Electron. Agricult., vol. 156, pp. 119–128, Jan. 2019.
arXiv:2311.03240. [51] P. R. Rothe and R. V. Kshirsagar, ‘‘Cotton leaf disease identification using
[30] S. Mustofa, M. M. H. Munna, Y. R. Emon, G. Rabbany, and M. T. Ahad, pattern recognition techniques,’’ in Proc. Int. Conf. Pervasive Comput.
‘‘A comprehensive review on plant leaf disease detection using deep (ICPC), Jan. 2015, pp. 1–6.
learning,’’ 2023, arXiv:2308.14087. [52] J. Abdulridha, Y. Ampatzidis, S. C. Kakarla, and P. Roberts, ‘‘Detection
of target spot and bacterial spot diseases in tomato using UAV-based
[31] L. Goel and J. Nagpal, ‘‘A systematic review of recent machine
and benchtop-based hyperspectral imaging techniques,’’ Precis. Agricult.,
learning techniques for plant disease identification and classifica-
vol. 21, no. 5, pp. 955–978, Oct. 2020.
tion,’’ IETE Tech. Rev., vol. 40, no. 3, pp. 423–439, Sep. 2022, doi:
10.1080/02564602.2022.2121772. [53] D. Zhang, X. Zhou, J. Zhang, Y. Lan, C. Xu, and D. Liang, ‘‘Detection of
rice sheath blight using an unmanned aerial system with high-resolution
[32] A. Bhargava and A. Bansal, ‘‘Fruits and vegetables quality evaluation
color and multispectral imaging,’’ PLoS ONE, vol. 13, no. 5, May 2018,
using computer vision: A review,’’ J. King Saud Univ.-Comput. Inf. Sci.,
Art. no. e0187470.
vol. 33, no. 3, pp. 243–257, Mar. 2021.
[54] K.-Y. Huang, ‘‘Application of artificial neural network for detecting
[33] S. P. Mohanty, D. P. Hughes, and M. Salathé, ‘‘Using deep learning for phalaenopsis seedling diseases using color and texture features,’’ Comput.
image-based plant disease detection,’’ Frontiers Plant Sci., vol. 7, p. 1419, Electron. Agricult., vol. 57, no. 1, pp. 3–11, May 2007.
Sep. 2016. [55] Q. Yao, Z. Guan, Y. Zhou, J. Tang, Y. Hu, and B. Yang, ‘‘Application of
[34] J. G. Arnal Barbedo, ‘‘Plant disease identification from individual lesions support vector machine for detecting rice diseases using shape and color
and spots using deep learning,’’ Biosyst. Eng., vol. 180, pp. 96–107, texture features,’’ in Proc. Int. Conf. Eng. Comput., May 2009, pp. 79–83.
Apr. 2019. [56] S. Kaur, S. Pandey, and S. Goel, ‘‘Semi-automatic leaf disease detection
[35] K. Bashir, M. Rehman, and M. Bari, ‘‘Detection and classification of rice and classification system for soybean culture,’’ IET Image Process.,
diseases: An automated approach using textural features,’’ Mehran Univ. vol. 12, no. 6, pp. 1038–1048, Jun. 2018.
Res. J. Eng. Technol., vol. 38, no. 1, pp. 239–250, Jan. 2019. [57] M. Francisco, F. Ribeiro, J. Metrôlho, and R. Dionísio, ‘‘Algorithms and
[36] A. Camargo and J. S. Smith, ‘‘An image-processing based algorithm models for automatic detection and classification of diseases and pests
to automatically identify plant disease visual symptoms,’’ Biosyst. Eng., in agricultural crops: A systematic review,’’ Appl. Sci., vol. 13, no. 8,
vol. 102, no. 1, pp. 9–21, Jan. 2009. p. 4720, Apr. 2023.
37462 VOLUME 12, 2024

[58] R. B. Dhanakotti et al., ‘‘Structural and magnetic properties of cobalt- [80] A. Rastogi, R. Arora, and S. Sharma, ‘‘Leaf disease detection and grading
doped iron oxide nanoparticles prepared by solution combustion method using computer vision technology & fuzzy logic,’’ in Proc. 2nd Int. Conf.
for biomedical applications,’’ Int. J. Nanomedicine, p. 189, Oct. 2015, Signal Process. Integr. Netw. (SPIN), Feb. 2015, pp. 500–505.
doi: 10.2147/ijn.s82210. [81] S. B. Jadhav and S. B. Patil, ‘‘Grading of soybean leaf disease based
[59] G. Dhingra, V. Kumar, and H. D. Joshi, ‘‘Study of digital image process- on segmented image using k-means clustering,’’ IAES Int. J. Artif. Intell.
ing techniques for leaf disease detection and classification,’’ Multimedia (IJ-AI), vol. 5, no. 1, p. 13, Mar. 2016.
Tools Appl., vol. 77, no. 15, pp. 19951–20000, Aug. 2018. [82] S. Zhang, Y. Zhu, Z. You, and X. Wu, ‘‘Fusion of superpixel, expectation
[60] A. Cruz, Y. Ampatzidis, R. Pierro, A. Materazzi, A. Panattoni, maximization and PHOG for recognizing cucumber diseases,’’ Comput.
L. De Bellis, and A. Luvisi, ‘‘Detection of grapevine yellows symp- Electron. Agricult., vol. 140, pp. 338–347, Aug. 2017.
toms in Vitis vinifera L. With artificial intelligence,’’ Comput. Electron. [83] S. S. Harakannanavar, J. M. Rudagi, V. I. Puranikmath, A. Siddiqua, and
Agricult., vol. 157, pp. 63–76, Feb. 2019. R. Pramodhini, ‘‘Plant leaf disease detection using computer vision and
[61] K. Vidyaraj and S. Priya, ‘‘Developing an algorithm for tomato leaf machine learning algorithms,’’ Global Transitions Proc., vol. 3, no. 1,
disease detection and classification,’’ Int. J. Innov. Res. Elect. Electron. pp. 305–310, Jun. 2022.
Instrum. Control Eng., vol. 3, no. 1, Feb. 2016. [84] S. Zhang, H. Wang, W. Huang, and Z. You, ‘‘Plant diseased leaf segmenta-
[62] S. Kai, L. Zhikun, S. Hang, and G. Chunhong, ‘‘A research of maize tion and recognition by fusion of superpixel, K-means and PHOG,’’ Optik,
disease image recognition of corn based on BP networks,’’ in Proc. 3rd vol. 157, pp. 866–872, Mar. 2018, doi: 10.1016/j.ijleo.2017.11.190.
Int. Conf. Measuring Technol. Mechatronics Autom., vol. 1, Jan. 2011, [85] M. Sachin, B. Jagtap, M. Shailesh, and M. Hambarde. Agricultural
pp. 246–249. Plant Leaf Disease Detection and Diagnosis Using Image Processing
[63] P. Chaudhary, A. K. Chaudhari, A. Cheeran, and S. Godara, ‘‘Color Based on Morphological Feature Extraction. Accessed: Aug. 9, 2023.
transform based approach for disease spot detection on plant leaf,’’ Int. [Online]. Available: https://www.iosrjournals.org/iosr-jvlsi/papers/vol4-
J. Comput. Sci. Telecommun., vol. 3, no. 6, pp. 65–70, 2012. issue5/Version-1/E04512430.pdf
[64] A. A. Joshi and B. D. Jadhav, ‘‘Monitoring and controlling rice diseases [86] X. Bai, X. Li, Z. Fu, X. Lv, and L. Zhang, ‘‘A fuzzy clustering
using image processing techniques,’’ in Proc. Int. Conf. Comput., Analyt- segmentation method based on neighborhood grayscale information
ics Secur. Trends (CAST), Dec. 2016, pp. 471–476. for defining cucumber leaf spot disease images,’’ Comput. Electron.
[65] P. Ganesan, G. Sajiv, and L. M. Leo, ‘‘CIELuv color space for identi- Agricult., vol. 136, pp. 157–165, Apr. 2017.
fication and segmentation of disease affected plant leaves using fuzzy [87] S. Kumar Sahu and M. Pandey, ‘‘An optimal hybrid multiclass SVM for
based approach,’’ in Proc. 3rd Int. Conf. Sci. Technol. Eng. Manag. plant leaf disease detection using spatial fuzzy C-means model,’’ Exp.
(ICONSTEM), Mar. 2017, pp. 889–894. Syst. Appl., vol. 214, Mar. 2023, Art. no. 118989.
[66] A. Meunkaewjinda, P. Kumsawat, K. Attakitmongcol, and A. Srikaew, [88] V. Singh and A. K. Misra, ‘‘Detection of unhealthy region of plant leaves
‘‘Grape leaf disease detection from color imagery using hybrid intelligent using image processing and genetic algorithm,’’ in Proc. Int. Conf. Adv.
system,’’ in Proc. 5th Int. Conf. Electr. Engineering/Electronics, Comput., Comput. Eng. Appl., Mar. 2015, pp. 1028–1032.
Telecommun. Inf. Technol., May 2008, pp. 513–516. [89] V. Singh and A. K. Misra, ‘‘Detection of plant leaf diseases using image
[67] A. Rao and S. B. Kulkarni, ‘‘A hybrid approach for plant leaf disease segmentation and soft computing techniques,’’ Inf. Process. Agricult.,
detection and classification using digital image processing methods,’’ Int. vol. 4, no. 1, pp. 41–49, Mar. 2017.
J. Electr. Eng. Educ., vol. 12, 2020, Art. no. 002072092095312. [90] S. S. Chouhan, U. P. Singh, and S. Jain, ‘‘Applications of computer vision
[68] A. Asfarian, Y. Herdiyeni, A. Rauf, and K. H. Mutaqin, ‘‘Paddy diseases in plant pathology: A survey,’’ Arch. Comput. Methods Eng., vol. 27, no. 2,
identification with texture analysis using fractal descriptors based on pp. 611–632, Apr. 2020.
Fourier spectrum,’’ in Proc. Int. Conf. Comput., Control, Informat. Appl. [91] U. Mokhtar, N. El Bendary, A. E. Hassenian, E. Emary, M. A. Mahmoud,
(ICINA), Nov. 2013, pp. 77–81. H. Hefny, and M. F. Tolba, ‘‘SVM-based detection of tomato leaves
[69] K. Thangadurai and K. Padmavathi, ‘‘Computer visionimage enhance- diseases,’’ in Advances in Intelligent Systems and Computing. Cham,
ment for plant leaves disease detection,’’ in Proc. World Congr. Comput. Switzerland: Springer, 2015, pp. 641–652.
Commun. Technol., Feb. 2014, pp. 173–175. [92] M. Bhagat and D. Kumar, ‘‘Efficient feature selection using BoWs and
[70] S. D. Khirade and A. B. Patil, ‘‘Plant disease detection using image SURF method for leaf disease identification,’’ Multimedia Tools Appl.,
processing,’’ in Proc. Int. Conf. Comput. Commun. Control Autom., vol. 82, no. 18, pp. 28187–28211, Feb. 2023, doi: 10.1007/s11042-023-
Feb. 2015, pp. 768–771. 14625-5.
[71] E. Albahari, H. Madzin, R. Wirza, O. K. Rahmat, and R. Mas, ‘‘Image [93] D. Al Bashish, M. Braik, and S. Bani-Ahmad, ‘‘Detection and clas-
enhancement techniques on plant leaf and seed disease detection,’’ Int. J. sification of leaf diseases using K-means-based segmentation and
Innov. Res. Comput. Commun. Eng., vol. 5, no. 1, pp. 109–116, 2019. neural-networks-based classification,’’ Inf. Technol. J., vol. 10, no. 2,
[72] S. Sladojevic, M. Arsenovic, A. Anderla, D. Culibrk, and D. Stefanovic, pp. 267–275, Jan. 2011.
‘‘Deep neural networks based recognition of plant diseases by leaf image [94] Y. Tian, C. Zhao, S. Lu, and X. Guo, ‘‘Multiple classifier combination for
classification,’’ Comput. Intell. Neurosci., vol. 2016, pp. 1–11, May 2016. recognition of wheat leaf diseases,’’ Intell. Autom. Soft Comput., vol. 17,
[73] P. Goncharov, G. Ososkov, A. Nechaevskiy, A. Uzhinskiy, and no. 5, pp. 519–529, Jan. 2011.
I. Nestsiarenia, ‘‘Disease detection on the plant leaves by deep learning,’’ [95] P. M. Mainkar, S. Ghorpade, and M. Adawadkar, ‘‘Plant leaf disease
in Advances in Neural Computation, Machine Learning, and Cognitive detection and classification using image processing techniques,’’ Int. J.
Research II. Cham, Switzerland: Springer, 2019, pp. 151–159. Innov. Emerg. Res. Eng., vol. 2, no. 4, pp. 139–144, 2015.
[74] J. G. A. Barbedo, ‘‘A review on the main challenges in automatic plant [96] M. Islam, A. Dinh, K. Wahid, and P. Bhowmik, ‘‘Detection of potato dis-
disease identification based on visible range images,’’ Biosystems Eng., eases using image segmentation and multiclass support vector machine,’’
vol. 144, pp. 52–60, Apr. 2016. in Proc. IEEE 30th Can. Conf. Electr. Comput. Eng. (CCECE), Apr. 2017,
[75] G. Anthonys and N. Wickramarachchi, ‘‘An image recognition system pp. 1–4.
for crop disease identification of paddy fields in Sri Lanka,’’ in Proc. Int. [97] M. Sharif, M. A. Khan, Z. Iqbal, M. F. Azam, M. I. U. Lali, and
Conf. Ind. Inf. Syst. (ICIIS), Sri Lanka, Dec. 2009, pp. 403–407. M. Y. Javed, ‘‘Detection and classification of citrus diseases in agricul-
[76] S. Bankar, A. Dube, P. Kadam, and S. Deokule, ‘‘Plant disease detection ture based on optimized weighted segmentation and feature selection,’’
techniques using Canny edge detection & color histogram in image pro- Comput. Electron. Agricult., vol. 150, pp. 220–234, Jul. 2018.
cessing,’’ Int. J. Comput. Sci. Inf. Technol., vol. 5, no. 2, pp. 1165–1168, [98] Y. Dandawate and R. Kokare, ‘‘An automated approach for classification
2014. of plant diseases towards development of futuristic decision support
[77] S. B. Jadhav, V. R. Udup, and S. B. Patil, ‘‘Soybean leaf disease detection system in Indian perspective,’’ in Proc. Int. Conf. Adv. Comput., Commun.
and severity measurement using multiclass SVM and KNN classifier,’’ Informat. (ICACCI), Aug. 2015, pp. 794–799.
Int. J. Electr. Comput. Eng. (IJECE), vol. 9, no. 5, p. 4077, Oct. 2019. [99] K. J. Mohan, M. Balasubramanian, and S. Palanivel, ‘‘Detection and
[78] J. Pang, Z.-Y. Bai, J.-C. Lai, and S.-K. Li, ‘‘Automatic segmentation recognition of diseases from paddy plant leaf images,’’ Int. J. Com-
of crop leaf spot disease images by integrating local threshold and put. Appl., vol. 144, no. 12, pp. 34–41, Jun. 2016, doi: 10.5120/
seeded region growing,’’ in Proc. Int. Conf. Image Anal. Signal Process., ijca2016910505.
Oct. 2011, pp. 590–594. [100] S. Ramesh, R. Hebbar, M. Niveditha, R. Pooja, N. Shashank, and
[79] V. Singh, S. Gupta, and S. Saini, ‘‘A methodological survey of image P. V. Vinod, ‘‘Plant disease detection using machine learning,’’ in Proc.
segmentation using soft computing techniques,’’ in Proc. Int. Conf. Adv. Int. Conf. Design Innov. 3Cs Compute Communicate Control (ICDIC),
Comput. Eng. Appl., Mar. 2015, pp. 419–422. Apr. 2018, pp. 41–45.
VOLUME 12, 2024 37463

[101] H. Waghmare, R. Kokare, and Y. Dandawate, ‘‘Detection and classifica- [121] S. Phadikar, J. Sil, and A. K. Das, ‘‘Rice diseases classification using
tion of diseases of grape plant using opposite colour local binary pattern feature selection and rule generation techniques,’’ Comput. Electron.
feature and machine learning for automated decision support system,’’ Agricult., vol. 90, pp. 76–85, Jan. 2013.
in Proc. 3rd Int. Conf. Signal Process. Integr. Netw. (SPIN), Feb. 2016, [122] R. R. Patil and S. Kumar, ‘‘Rice-fusion: A multimodality data
pp. 513–518. fusion framework for rice disease diagnosis,’’ IEEE Access, vol. 10,
[102] K. Singh, S. Kumar, and P. Kaur, ‘‘Automatic detection of rust disease pp. 5207–5222, 2022.
of lentil by machine learning system using microscopic images,’’ Int. J. [123] N. Sengar, M. K. Dutta, and C. M. Travieso, ‘‘Computer vision based
Electr. Comput. Eng. (IJECE), vol. 9, no. 1, p. 660, Feb. 2019. technique for identification and quantification of powdery mildew dis-
[103] V. A. Gulhane and A. A. Gurjar, ‘‘Detection of diseases on cot- ease in cherry leaves,’’ Computing, vol. 100, no. 11, pp. 1189–1201,
ton leaves and its possible diagnosis,’’ Int. J. Image Process., vol. 5, Nov. 2018.
no. 5, pp. 590–598, 2011. [Online]. Available: https://www.cscjournals. [124] O. Y. Al-Jarrah, A. Siddiqui, M. Elsalamouny, P. D. Yoo, S. Muhaidat,
org/library/manuscriptinfo.php?mc=IJIP-478 and K. Kim, ‘‘Machine-Learning-Based feature selection techniques for
[104] S. Prasad, P. Kumar, R. Hazra, and A. Kumar, ‘‘Plant leaf disease large-scale network intrusion detection,’’ in Proc. IEEE 34th Int. Conf.
detection using Gabor wavelet transform,’’ in Swarm, Evolutionary, and Distrib. Comput. Syst. Workshops (ICDCSW), Jun. 2014, pp. 177–181.
Memetic Computing. Berlin, Germany: Springer, 2012, pp. 372–379. [125] V. K. Shrivastava and M. K. Pradhan, ‘‘Rice plant disease classifi-
[105] A. Kulkarni, ‘‘Applying image processing technique to detect plant dis- cation using color features: A machine learning paradigm,’’ J. Plant
eases,’’ Int. J. Modern Eng. Res., vol. 2, no. 5, pp. 3661–3664, 2012. Pathol., vol. 103, no. 1, pp. 17–26, Oct. 2020, doi: 10.1007/s42161-020-
[106] P. Jolly and S. Raman, ‘‘Analyzing surface defects in apples using Gabor 00683-3.
features,’’ in Proc. 12th Int. Conf. Signal-Image Technol. Internet-Based [126] A. K. Rath and J. K. Meher, ‘‘Disease detection in infected plant leaf
Syst. (SITIS), Nov. 2016, pp. 178–185. by computational method,’’ Arch. Phytopathol. Plant Protection, vol. 52,
[107] H. Sabrol and S. Kumar, ‘‘Fuzzy and neural network based tomato nos. 19–20, pp. 1348–1358, Dec. 2019.
plant disease classification using natural outdoor images,’’ Indian J. Sci. [127] J. Luo, S. Geng, C. Xiu, D. Song, and T. Dong, ‘‘A curvelet-SC recog-
Technol., vol. 9, no. 44, pp. 1–8, Nov. 2016. nition method for maize disease,’’ J. Electr. Comput. Eng., vol. 2015,
[108] S. M. Kiran and D. N. Chandrappa, ‘‘Plant disease identification using pp. 1–8, Jan. 2015.
discrete wavelet transforms and SVM,’’ J. Univ. Shanghai Sci. Technol., [128] S. Gharge and P. Singh, ‘‘Image processing for soybean disease clas-
vol. 23, no. 6, pp. 108–114, 2021. [Online]. Available: https://jusst. sification and severity estimation,’’ in Emerging Research in Comput-
org/wp-content/uploads/2021/06/Plant-Disease-Identification-Using- ing, Information, Communication and Applications. New Delhi, India:
Discrete-Wavelet-Transforms-and-SVM-1.pdf Springer, 2016, pp. 493–500.
[109] N. Dalal and B. Triggs, ‘‘Histograms of oriented gradients for human [129] A. Kaya, A. S. Keceli, C. Catal, H. Y. Yalic, H. Temucin, and
detection,’’ in Proc. IEEE Comput. Soc. Conf. Comput. Vis. Pattern B. Tekinerdogan, ‘‘Analysis of transfer learning for deep neural network
Recognit., Jun. 2005, pp. 886–893. [Online]. Available: http://vision. based plant classification models,’’ Comput. Electron. Agricult., vol. 158,
stanford.edu/teaching/cs231b_spring1213/papers/CVPR05_ pp. 20–29, Mar. 2019.
DalalTriggs.pdf [130] A. Sivasangari and K. Priya, ‘‘Cotton leaf disease detection and recov-
[110] Y. Bai, L. Guo, L. Jin, and Q. Huang, ‘‘A novel feature extraction ery using genetic algorithm,’’ Int. J. Eng. Res. Gen. Sci., vol. 117,
method using pyramid histogram of orientation gradients for smile recog- pp. 119–123, 2014. [Online]. Available: https://www.acadpubl.eu/jsi/
nition,’’ in Proc. 16th IEEE Int. Conf. Image Process. (ICIP), Nov. 2009, 2017-117-20-22/articles/22/23.pdf
pp. 3305–3308. [131] L. Hallau, M. Neumann, B. Klatt, B. Kleinhenz, T. Klein, C. Kuhn,
[111] S. S. Sannakki, V. S. Rajpurohit, V. B. Nargund, and P. Kulkarni, ‘‘Diag- M. Röhrig, C. Bauckhage, K. Kersting, A. Mahlein, U. Steiner, and
nosis and classification of grape leaf diseases using neural networks,’’ E. Oerke, ‘‘Automated identification of sugar beet diseases using smart-
in Proc. 4th Int. Conf. Comput., Commun. Netw. Technol. (ICCCNT), phones,’’ Plant Pathol., vol. 67, no. 2, pp. 399–410, Feb. 2018.
Jul. 2013, pp. 1–5. [132] I. Ahmed and P. K. Yadav, ‘‘Plant disease detection using machine
[112] R. D. L. Pires, D. N. Gonçalves, J. P. M. Oruě, W. E. S. Kanashiro, learning approaches,’’ Exp. Syst., vol. 40, no. 5, Oct. 2022, doi:
J. F. Rodrigues, B. B. Machado, and W. N. Gonçalves, ‘‘Local descriptors 10.1111/exsy.13136.
for soybean disease recognition,’’ Comput. Electron. Agricult., vol. 125, [133] B. J. Samajpati and S. D. Degadwala, ‘‘Hybrid approach for apple
pp. 48–55, Jul. 2016. fruit diseases detection and classification using random forest classi-
[113] B. S. Kusumo, A. Heryana, O. Mahendra, and H. F. Pardede, ‘‘Machine fier,’’ in Proc. Int. Conf. Commun. Signal Process. (ICCSP), Apr. 2016,
learning-based for automatic detection of corn-plant diseases using pp. 1015–1019.
image processing,’’ in Proc. Int. Conf. Comput., Control, Informat. Appl. [134] S. M. Javidan, A. Banakar, K. A. Vakilian, and Y. Ampatzidis, ‘‘Diagnosis
(IC3INA), Nov. 2018, pp. 93–97. of grape leaf diseases using automatic K-means clustering and machine
[114] A. S. Zamani, L. Anand, K. P. Rane, P. Prabhu, A. M. Buttar, learning,’’ Smart Agricult. Technol., vol. 3, Feb. 2023, Art. no. 100081,
H. Pallathadka, A. Raghuvanshi, and B. N. Dugbakie, ‘‘Performance of doi: 10.1016/j.atech.2022.100081.
machine learning and image processing in plant leaf disease detection,’’ [135] M. Shantkumari and S. V. Uma, ‘‘Grape leaf image classification based
J. Food Qual., vol. 2022, pp. 1–7, Apr. 2022. on machine learning technique for accurate leaf disease detection,’’
[115] P. Revathi and M. Hemalatha, ‘‘Cotton leaf spot diseases detection Multimedia Tools Appl., vol. 82, no. 1, pp. 1477–1487, Apr. 2022, doi:
utilizing feature selection with skew divergence method,’’ Int. J. Sci. 10.1007/s11042-022-12976-z.
Eng. Technol., vol. 3, no. 1, pp. 22–30, 2014. [Online]. Available: [136] P. Patil, N. Yaligar, and S. M. Meena, ‘‘Comparision of performance
http://www.ijset.com/publication/v3/005.pdf of classifiers—SVM, RF and ANN in potato blight disease detection
[116] M. Ramakrishnan, ‘‘Groundnut leaf disease detection and classification using leaf images,’’ in Proc. IEEE Int. Conf. Comput. Intell. Comput. Res.
by using back probagation algorithm,’’ in Proc. Int. Conf. Commun. (ICCIC), Dec. 2017, pp. 1–5.
Signal Process. (ICCSP), Apr. 2015, pp. 0964–0968. [137] S. Verma, A. Chug, and A. P. Singh, ‘‘Exploring capsule networks for
[117] A. Caglayan, O. Guclu, and A. B. Can, ‘‘A plant recognition approach disease classification in plants,’’ J. Statist. Manag. Syst., vol. 23, no. 2,
using shape and color features in leaf images,’’ in Image Analysis and pp. 307–315, Feb. 2020.
Processing—ICIAP. Berlin, Germany: Springer, 2013, pp. 161–170. [138] A. Krishnakumar and A. Narayanan, ‘‘A system for plant disease clas-
[118] T. Munisami, M. Ramsurn, S. Kishnah, and S. Pudaruth, ‘‘Plant sification and severity estimation using machine learning techniques,’’
leaf recognition using shape features and colour histogram with in Proc. Int. Conf. ISMAC Comput. Vis. Bio-Eng. (Lecture Notes in
K-nearest neighbour classifiers,’’ Proc. Comput. Sci., vol. 58, Computational Vision and Biomechanics), Jan. 2019, pp. 447–457.
pp. 740–747, Jan. 2015. [139] S.-E.-A. Raza, G. Prince, J. P. Clarkson, and N. M. Rajpoot, ‘‘Automatic
[119] A. Camargo and J. S. Smith, ‘‘Image pattern classification for the identi- detection of diseased tomato plants using thermal and stereo visible light
fication of disease causing agents in plants,’’ Comput. Electron. Agricult., images,’’ PLoS ONE, vol. 10, no. 4, Apr. 2015, Art. no. e0123262.
vol. 66, no. 2, pp. 121–125, May 2009. [140] B. A. M. Ashqar and S. S. Abu-Naser. Image-Based Tomato Leaves
[120] H. Wang, G. Li, Z. Ma, and X. Li, ‘‘Application of neural networks to Diseases Detection Using Deep Learning. Accessed: Aug. 10, 2023.
image recognition of plant diseases,’’ in Proc. Int. Conf. Syst. Informat. [Online]. Available: http://dstore.alazhar.edu.ps/xDLui/bitstream/handle/
(ICSAI), May 2012, pp. 2159–2164. 123456789/278/ASHITL-3v1.pdf?sequence=1&isAllowed=y
37464 VOLUME 12, 2024

[141] S. Jasrotia, J. Yadav, N. Rajpal, M. Arora, and J. Chaudhary, ‘‘Convo- [165] D. Ho, E. Liang, I. Stoica, P. Abbeel, and X. Chen, ‘‘Population based aug-
lutional neural network based maize plant disease identification,’’ Proc. mentation: Efficient learning of augmentation policy schedules,’’ 2019,
Comput. Sci., vol. 218, pp. 1712–1721, Jan. 2023. arXiv:1905.05393.
[142] R. Sujatha, J. M. Chatterjee, N. Jhanjhi, and S. N. Brohi, ‘‘Performance [166] E. D. Cubuk, B. Zoph, J. Shlens, and Q. V. Le, ‘‘RandAugment: Prac-
of deep learning vs machine learning in plant leaf disease detection,’’ tical automated data augmentation with a reduced search space,’’ 2019,
Microprocess. Microsyst., vol. 80, Feb. 2021, Art. no. 103615. arXiv:1909.13719.
[143] S. U. Rahman, F. Alam, N. Ahmad, and S. Arshad, ‘‘Image processing [167] I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley,
based system for the detection, identification and treatment of tomato S. Ozair, A. Courville, and Y. Bengio, ‘‘Generative adversarial nets,’’ in
leaf diseases,’’ Multimedia Tools Appl., vol. 82, no. 6, pp. 9431–9445, Proc. Adv. Neural Inf. Process. Syst., 2014, pp. 1–9. [Online]. Available:
Sep. 2022. https://proceedings.neurips.cc/paper_files/paper/2014/file/5ca3e9b122f6
[144] A. Bhatia, A. Chug, and A. Prakash Singh, ‘‘Application of extreme learn- 1f8f06494c97b1afccf3-Paper.pdf
ing machine in plant disease prediction for highly imbalanced dataset,’’ [168] B. Liu, Y. Zhang, D. He, and Y. Li, ‘‘Identification of apple leaf diseases
J. Statist. Manag. Syst., vol. 23, no. 6, pp. 1059–1068, Jul. 2020. based on deep convolutional neural networks,’’ Symmetry, vol. 10, no. 1,
[145] A. Bhatia, A. Chug, and A. P. Singh, ‘‘Hybrid SVM-LR classifier for p. 11, Dec. 2017.
powdery mildew disease prediction in tomato plant,’’ in Proc. 7th Int. [169] R. Chen, H. Qi, Y. Liang, and M. Yang, ‘‘Identification of plant leaf dis-
Conf. Signal Process. Integr. Netw., Feb. 2020, pp. 218–223. [Online]. eases by deep learning based on channel attention and channel pruning,’’
Available: https://ieeexplore.ieee.org/document/9071202 Frontiers Plant Sci., vol. 13, Nov. 2022.
[146] W. Mcculloch and W. Pitts, ‘‘A logical calculus of the ideas immanent in [170] H. Nazki, S. Yoon, A. Fuentes, and D. S. Park, ‘‘Unsupervised image
nervous activity,’’ Bull. Math. Biol., vol. 52, nos. 1–2, pp. 99–115, 1990. translation using adversarial networks for improved plant disease recog-
[147] M. Dyrmann, H. Karstoft, and H. S. Midtiby, ‘‘Plant species classification nition,’’ Comput. Electron. Agricult., vol. 168, Jan. 2020, Art. no. 105117.
using deep convolutional neural network,’’ Biosystems Eng., vol. 151, [171] Y. Tian, G. Yang, Z. Wang, E. Li, and Z. Liang, ‘‘Detection of apple
pp. 72–80, Nov. 2016. lesions in orchards based on deep learning methods of CycleGAN and
[148] N. Kussul, M. Lavreniuk, S. Skakun, and A. Shelestov, ‘‘Deep learning YOLOV3-dense,’’ J. Sensors, vol. 2019, pp. 1–13, Apr. 2019.
classification of land cover and crop types using remote sensing data,’’ [172] Q. Wu, Y. Chen, and J. Meng, ‘‘DCGAN-based data augmenta-
IEEE Geosci. Remote Sens. Lett., vol. 14, no. 5, pp. 778–782, May 2017. tion for tomato leaf disease identification,’’ IEEE Access, vol. 8,
[149] R. Wason, ‘‘Deep learning: Evolution and expansion,’’ Cognit. Syst. Res., pp. 98716–98728, 2020.
vol. 52, pp. 701–708, Dec. 2018. [173] B. Liu, C. Tan, S. Li, J. He, and H. Wang, ‘‘A data augmentation method
[150] G. Geetharamani and A. Pandian, ‘‘Identification of plant leaf diseases based on generative adversarial networks for grape leaf disease identifi-
using a nine-layer deep convolutional neural network,’’ Comput. Electr. cation,’’ IEEE Access, vol. 8, pp. 102188–102198, 2020.
Eng., vol. 76, pp. 323–338, Jun. 2019. [174] L. Ale, A. Sheta, L. Li, Y. Wang, and N. Zhang, ‘‘Deep learning based
[151] G. Huang, Z. Liu, L. Van Der Maaten, and K. Q. Weinberger, ‘‘Densely plant disease detection for smart agriculture,’’ in Proc. IEEE Globecom
connected convolutional networks,’’ in Proc. IEEE Conf. Comput. Vis. Workshops (GC Wkshps), Dec. 2019, pp. 1–6.
Pattern Recognit. (CVPR), Jul. 2017, pp. 2261–2269. [175] A. Darwish, D. Ezzat, and A. E. Hassanien, ‘‘An optimized model based
[152] E. C. Too, L. Yujian, S. Njuki, and L. Yingchun, ‘‘A comparative study on convolutional neural networks and orthogonal learning particle swarm
of fine-tuning deep learning models for plant disease identification,’’ optimization algorithm for plant diseases diagnosis,’’ Swarm Evol. Com-
Comput. Electron. Agricult., vol. 161, pp. 272–279, Jun. 2019. put., vol. 52, Feb. 2020, Art. no. 100616.
[153] K. He, X. Zhang, S. Ren, and J. Sun, ‘‘Identity mappings in deep residual [176] R. S. Yuwana, E. Suryawati, V. Zilvan, A. Ramdan, H. F. Pardede,
networks,’’ in Computer Vision—ECCV. Cham, Switzerland: Springer, and F. Fauziah, ‘‘Multi-condition training on deep convolutional neural
2016, pp. 630–645. networks for robust plant diseases detection,’’ in Proc. Int. Conf. Comput.,
[154] Z. Gao, Z. Luo, W. Zhang, Z. Lv, and Y. Xu, ‘‘Deep learning applica- Control, Informat. Appl. (ICINA), Oct. 2019, pp. 30–35.
tion in plant stress imaging: A review,’’ AgriEngineering, vol. 2, no. 3, [177] J. Chen, J. Chen, D. Zhang, Y. Sun, and Y. A. Nanehkaran, ‘‘Using deep
pp. 430–446, Jul. 2020. transfer learning for image-based plant disease identification,’’ Comput.
[155] K. He, X. Zhang, S. Ren, and J. Sun, ‘‘Deep residual learning for image Electron. Agricult., vol. 173, Jun. 2020, Art. no. 105393.
recognition,’’ 2015, arXiv:1512.03385. [178] S. Mishra, R. Sachan, and D. Rajpal, ‘‘Deep convolutional neural network
[156] A. G. Howard, M. Zhu, B. Chen, D. Kalenichenko, W. Wang, T. Weyand, based detection system for real-time corn plant disease recognition,’’
M. Andreetto, and H. Adam, ‘‘MobileNets: Efficient convolutional neural Proc. Comput. Sci., vol. 167, pp. 2003–2010, Jan. 2020.
networks for mobile vision applications,’’ 2017, arXiv:1704.04861. [179] N. K. Trivedi, V. Gautam, A. Anand, H. M. Aljahdali, S. G. Villar,
[157] A. Ahmad, D. Saraswat, and A. El Gamal, ‘‘A survey on using deep D. Anand, N. Goyal, and S. Kadry, ‘‘Early detection and classification
learning techniques for plant disease diagnosis and recommendations of tomato leaf disease using high-performance deep neural network,’’
for development of appropriate tools,’’ Smart Agricult. Technol., vol. 3, Sensors, vol. 21, no. 23, p. 7987, Nov. 2021.
Feb. 2023, Art. no. 100083, doi: 10.1016/j.atech.2022.100083. [180] P. Deshwal, K. Sharma, and M. S. Moudgil, ‘‘Plant leaf disease detection
[158] M. Türkoglu and D. Hanbay, ‘‘Plant disease and pest detection using deep using machine learning,’’ Int. J. Res. Appl. Sci. Eng. Technol., vol. 11,
learning-based features,’’ TURKISH J. Electr. Eng. Comput. Sci., vol. 27, no. 5, pp. 5928–5932, May 2023.
no. 3, pp. 1636–1651, May 2019. [181] M. A. Jasim and J. M. Al-Tuwaijari, ‘‘Plant leaf diseases detection and
[159] M. Brahimi, S. Mahmoudi, K. Boukhalfa, and A. Moussaoui, ‘‘Deep classification using image processing and deep learning techniques,’’
interpretable architecture for plant diseases classification,’’ in Proc. Sig- in Proc. Int. Conf. Comput. Sci. Softw. Eng. (CSASE), Apr. 2020,
nal Process., Algorithms, Architectures, Arrangements, Appl., Poznan, pp. 259–265.
Poland, Sep. 2019, pp. 111–116. [182] A. V. Panchal, S. C. Patel, K. Bagyalakshmi, P. Kumar, I. R. Khan, and
[160] M. Mehdipour Ghazi, B. Yanikoglu, and E. Aptoula, ‘‘Plant identification M. Soni, ‘‘Image-based plant diseases detection using deep learning,’’
using deep neural networks via optimization of transfer learning parame- Mater. Today, Proc., vol. 80, pp. 3500–3506, Jan. 2023.
ters,’’ Neurocomputing, vol. 235, pp. 228–235, Apr. 2017. [183] A. Chug, A. Bhatia, A. P. Singh, and D. Singh, ‘‘A novel frame-
[161] A. Ramcharan, K. Baranowski, P. McCloskey, B. Ahmed, J. Legg, and work for image-based plant disease detection using hybrid deep
D. P. Hughes, ‘‘Deep learning for image-based cassava disease detec- learning approach,’’ Soft Comput., vol. 27, no. 18, pp. 13613–13638,
tion,’’ Frontiers Plant Sci., vol. 8, p. 1852, Oct. 2017. Jun. 2022.
[162] Z. Lin, S. Mu, A. Shi, C. Pang, and X. Sun, ‘‘A novel method of maize [184] Q. Wang, F. Qi, M. Sun, J. Qu, and J. Xue, ‘‘Identification of tomato
leaf disease image identification based on a multichannel convolutional disease types and detection of infected areas based on deep convolutional
neural network,’’ Trans. ASABE, vol. 61, no. 5, pp. 1461–1474, 2018. neural networks and object detection techniques,’’ Comput. Intell. Neu-
[163] D. Hendrycks, N. Mu, E. D. Cubuk, B. Zoph, J. Gilmer, and rosci., vol. 2019, pp. 1–15, Dec. 2019.
B. Lakshminarayanan, ‘‘AugMix: A simple data processing method to [185] A. Picon, M. Seitz, A. Alvarez-Gila, P. Mohnke, A. Ortiz-Barredo,
improve robustness and uncertainty,’’ 2019, arXiv:1912.02781. and J. Echazarra, ‘‘Crop conditional convolutional neural networks for
[164] S. Yun, D. Han, S. Joon Oh, S. Chun, J. Choe, and Y. Yoo, ‘‘CutMix: Reg- massive multi-crop plant disease classification over cell phone acquired
ularization strategy to train strong classifiers with localizable features,’’ images taken on real field conditions,’’ Comput. Electron. Agricult.,
2019, arXiv:1905.04899. vol. 167, Dec. 2019, Art. no. 105093.
VOLUME 12, 2024 37465

[186] T. R. Gadekallu, D. S. Rajput, M. P. K. Reddy, K. Lakshmanna, [207] M. A. Khan, A. Alqahtani, A. Khan, S. Alsubai, A. Binbusayyis,
S. Bhattacharya, S. Singh, A. Jolfaei, and M. Alazab, ‘‘A novel PCA- M. M. I. Ch, H.-S. Yong, and J. Cha, ‘‘Cucumber leaf diseases recogni-
whale optimization-based deep neural network model for classification of tion using multi level deep entropy-ELM feature selection,’’ Appl. Sci.,
tomato plant diseases using GPU,’’ J. Real-Time Image Process., vol. 18, vol. 12, no. 2, p. 593, Jan. 2022, doi: 10.3390/app12020593.
no. 4, pp. 1383–1396, Aug. 2021. [208] H. Yu, X. Cheng, C. Chen, A. A. Heidari, J. Liu, Z. Cai, and H. Chen,
[187] M. Kaushik, P. Prakash, R. Ajay, and S. Veni, ‘‘Tomato leaf disease ‘‘Apple leaf disease recognition method with improved residual net-
detection using convolutional neural network with data augmentation,’’ work,’’ Multimedia Tools Appl., vol. 81, no. 6, pp. 7759–7782, Jan. 2022,
in Proc. 5th Int. Conf. Commun. Electron. Syst. (ICCES), Coimbatore, doi: 10.1007/s11042-022-11915-2.
India, Jun. 2020, pp. 1125–1132. [209] A. I. Khan, S. M. K. Quadri, S. Banday, and J. Latief Shah, ‘‘Deep diagno-
[188] S. Verma, A. Chug, and A. P. Singh, ‘‘Application of convolutional neural sis: A real-time apple leaf disease detection system based on deep learn-
networks for evaluation of disease severity in tomato plant,’’ J. Discrete ing,’’ Comput. Electron. Agricult., vol. 198, Jul. 2022, Art. no. 107093,
Math. Sci. Cryptogr., vol. 23, no. 1, pp. 273–282, Jan. 2020. doi: 10.1016/j.compag.2022.107093.
[189] P. Wspanialy and M. Moussa, ‘‘A detection and severity estimation sys- [210] Researchgate.net. Accessed: Aug. 10, 2023. [Online]. Available:
tem for generic diseases of tomato greenhouse plants,’’ Comput. Electron. https://www.researchgate.net/publication/366239200_Leaf_disease_
Agricult., vol. 178, Nov. 2020, Art. no. 105701. identification_and_classification_using_optimized_deep_learning
[190] R. Karthik, M. Hariharan, S. Anand, P. Mathikshara, A. Johnson, and [211] M. Shoaib, B. Shah, S. Ei-Sappagh, A. Ali, A. Ullah, F. Alenezi,
R. Menaka, ‘‘Attention embedded residual CNN for disease detection in T. Gechev, T. Hussain, and F. Ali, ‘‘An advanced deep learning models-
tomato leaves,’’ Appl. Soft Comput., vol. 86, Jan. 2020, Art. no. 105933. based plant disease detection: A review of recent research,’’ Frontiers
[191] I. Ahmad, M. Hamid, S. Yousaf, S. T. Shah, and M. O. Ahmad, ‘‘Opti- Plant Sci., vol. 14, Mar. 2023, Art. no. 1158933.
mizing pretrained convolutional neural networks for tomato leaf disease [212] P. Singh, P. Singh, U. Farooq, S. S. Khurana, J. K. Verma, and M. Kumar,
detection,’’ Complexity, vol. 2020, pp. 1–6, Sep. 2020. ‘‘CottonLeafNet: Cotton plant leaf disease detection using deep neural
[192] S. H. Lee, H. Goëau, P. Bonnet, and A. Joly, ‘‘New perspectives on networks,’’ Multimedia Tools Appl., vol. 82, no. 24, pp. 37151–37176,
plant disease characterization based on deep learning,’’ Comput. Electron. Oct. 2023.
Agricult., vol. 170, Mar. 2020, Art. no. 105220. [213] X-Mol X-Mol.Net. Accessed: Aug. 10, 2023. [Online]. Available:
[193] Ü. Atila, M. Uçar, K. Akyol, and E. Uçar, ‘‘Plant leaf disease classification https://www.x-mol.net/paper/article/1626716389179998208
using EfficientNet deep learning model,’’ Ecological Informat., vol. 61, [214] T. Daniya and S. Vigneshwari, ‘‘Rice plant leaf disease detection and
Mar. 2021, Art. no. 101182. classification using optimization enabled deep learning,’’ J. Environ.
[194] M. Agarwal, A. Singh, S. Arjaria, A. Sinha, and S. Gupta, ‘‘ToLeD: Inform., vol. 42, p. 25, Sep. 2023.
Tomato leaf disease detection using convolution neural network,’’ Proc. [215] R. Arumuga Arun and S. Umamaheswari, ‘‘Effective multi-crop dis-
Comput. Sci., vol. 167, pp. 293–301, Jan. 2020. ease detection using pruned complete concatenated deep learning
[195] L. C. Ngugi, M. Abdelwahab, and M. Abo-Zahhad, ‘‘Tomato leaf seg- model,’’ Exp. Syst. Appl., vol. 213, Mar. 2023, Art. no. 118905, doi:
mentation algorithms for mobile phone applications using deep learning,’’ 10.1016/j.eswa.2022.118905.
Comput. Electron. Agricult., vol. 178, Nov. 2020, Art. no. 105788. [216] A. Pal and V. Kumar, ‘‘AgriDet: Plant leaf disease severity classifica-
[196] J. Chen, J. Chen, D. Zhang, Y. A. Nanehkaran, and Y. Sun, ‘‘A cognitive tion using agriculture detection framework,’’ Eng. Appl. Artif. Intell.,
vision method for the detection of plant disease images,’’ Mach. Vis. vol. 119, Mar. 2023, Art. no. 105754, doi: 10.1016/j.engappai.2022.
Appl., vol. 32, no. 1, pp. 1–18, Jan. 2021. 105754.
[197] V. G. Krishnan, J. Deepa, P. V. Rao, V. Divya, and S. Kaviarasan, [217] S. R. G. Reddy, G. P. S. Varma, and R. L. Davuluri, ‘‘Resnet-based
‘‘An automated segmentation and classification model for banana leaf modified red deer optimization with DLCNN classifier for plant dis-
disease detection,’’ J. Appl. Biol. Biotechnol., vol. 10, pp. 213–220, ease identification and classification,’’ Comput. Electr. Eng., vol. 105,
Oct. 2022. Jan. 2023, Art. no. 108492, doi: 10.1016/j.compeleceng.2022.108492.
[198] R. K. Singh, A. Tiwari, and R. K. Gupta, ‘‘Deep transfer modeling [218] Y. Kaya and E. Gürsoy, ‘‘A novel multi-head CNN design to identify plant
for classification of maize plant leaf disease,’’ Multimedia Tools Appl., diseases using the fusion of RGB images,’’ Ecological Informat., vol. 75,
vol. 81, no. 5, pp. 6051–6067, Feb. 2022. Jul. 2023, Art. no. 101998, doi: 10.1016/j.ecoinf.2023.101998.
[199] A. Elaraby, W. Hamdy, and M. Alruwaili, ‘‘Optimization of deep learn- [219] H.-T. Thai, K.-H. Le, and N. L.-T. Nguyen, ‘‘FormerLeaf: An effi-
ing model for plant disease detection using particle swarm optimizer,’’ cient vision transformer for cassava leaf disease detection,’’ Com-
Comput., Mater. Continua, vol. 71, no. 2, pp. 4019–4031, 2022. put. Electron. Agricult., vol. 204, Jan. 2023, Art. no. 107518, doi:
[200] S. Ashwinkumar, S. Rajagopal, V. Manimaran, and B. Jegajothi, ‘‘Auto- 10.1016/j.compag.2022.107518.
mated plant leaf disease detection and classification using optimal [220] P. S. Thakur, T. Sheorey, and A. Ojha, ‘‘VGG-ICNN: A lightweight CNN
MobileNet based convolutional neural networks,’’ Mater. Today, Proc., model for crop disease identification,’’ Multimedia Tools Appl., vol. 82,
vol. 51, pp. 480–487, Jan. 2022. no. 1, pp. 497–520, Jun. 2022, doi: 10.1007/s11042-022-13144-z.
[201] J. A. Pandian, K. Kanchanadevi, V. D. Kumar, E. Jasinska, R. Gono, [221] P. Baranowski, M. Jedryczka, W. Mazurek, D. Babula-Skowronska,
Z. Leonowicz, and M. Jasinski, ‘‘A five convolutional layer deep con- A. Siedliska, and J. Kaczmarek, ‘‘Hyperspectral and thermal imaging of
volutional neural network for plant leaf disease detection,’’ Electronics, oilseed rape (Brassica napus) response to fungal species of the genus
vol. 11, no. 8, p. 1266, Apr. 2022. Alternaria,’’ PLoS ONE, vol. 10, no. 3, Mar. 2015, Art. no. e0122913.
[202] R. Mahum, H. Munir, Z.-U.-N. Mughal, M. Awais, F. S. Khan, [222] P. Ghamisi, J. Plaza, Y. Chen, J. Li, and A. J. Plaza, ‘‘Advanced spectral
M. Saqlain, S. Mahamad, and I. Tlili, ‘‘A novel framework for potato leaf classifiers for hyperspectral images: A review,’’ IEEE Geosci. Remote
disease detection using an efficient deep learning model,’’ Hum. Ecol. Sens. Mag., vol. 5, no. 1, pp. 8–32, Mar. 2017.
Risk Assessment, Int. J., vol. 29, no. 2, pp. 303–326, Apr. 2022, doi: [223] D. Moshou, C. Bravo, R. Oberti, J. West, L. Bodria, A. McCartney,
10.1080/10807039.2022.2064814. and H. Ramon, ‘‘Plant disease detection based on data fusion of hyper-
[203] A. Haridasan, J. Thomas, and E. D. Raj, ‘‘Deep learning system for paddy spectral and multi-spectral fluorescence imaging using Kohonen maps,’’
plant disease detection and classification,’’ Environ. Monitor. Assessment, Real-Time Imag., vol. 11, no. 2, pp. 75–83, Apr. 2005.
vol. 195, no. 1, p. 120, Nov. 2022, doi: 10.1007/s10661-022-10656-x. [224] A.-K. Mahlein, U. Steiner, C. Hillnhütter, H.-W. Dehne, and E.-C. Oerke,
[204] Y. M. A. Algani, O. J. M. Caro, L. M. R. Bravo, C. Kaur, ‘‘Hyperspectral imaging for small-scale analysis of symptoms caused by
M. S. Al Ansari, and B. K. Bala, ‘‘Leaf disease identification and classifi- different sugar beet diseases,’’ Plant Methods, vol. 8, no. 1, p. 3, 2012.
cation using optimized deep learning,’’ Meas., Sensors, vol. 25, Feb. 2023, [225] L. S. Huang, J. L. Zhao, D. Y. Zhang, L. Yuan, Y. Y. Dong, and
Art. no. 100643, doi: 10.1016/j.measen.2022.100643. J. C. Zhang, ‘‘Identifying and mapping stripe rust in winter wheat using
[205] X. Huang, A. Chen, G. Zhou, X. Zhang, J. Wang, N. Peng, N. Yan, and multi-temporal airborne hyperspectral images,’’ Int. J. Agricult. Biol.,
C. Jiang, ‘‘Tomato leaf disease detection system based on FC-SNDPN,’’ vol. 14, pp. 697–704, 2012.
Multimedia Tools Appl., vol. 82, no. 2, pp. 2121–2144, Jun. 2022, doi: [226] K. Nagasubramanian, S. Jones, A. K. Singh, A. Singh,
10.1007/s11042-021-11790-3. B. Ganapathysubramanian, and S. Sarkar, ‘‘Explaining hyperspectral
[206] S. Yu, L. Xie, and Q. Huang, ‘‘Inception convolutional vision transform- imaging based plant disease identification: 3D CNN and saliency maps,’’
ers for plant disease identification,’’ Internet Things, vol. 21, Apr. 2023, 2018, arXiv:1804.08831. [Online]. Available: http://www.interpretable-
Art. no. 100650, doi: 10.1016/j.iot.2022.100650. ml.org/nips2017workshop/papers/16.pdf
37466 VOLUME 12, 2024

[227] E. Alisaac, J. Behmann, M. T. Kuska, H.-W. Dehne, and A.-K. Mahlein, [246] M. Wahabzada, A.-K. Mahlein, C. Bauckhage, U. Steiner, E.-C. Oerke,
‘‘Hyperspectral quantification of wheat resistance to fusarium head and K. Kersting, ‘‘Metro maps of plant disease dynamics—Automated
blight: Comparison of two fusarium species,’’ Eur. J. Plant Pathol., mining of differences using hyperspectral images,’’ PLoS ONE, vol. 10,
vol. 152, no. 4, pp. 869–884, Dec. 2018. no. 1, Jan. 2015, Art. no. e0116902.
[228] J. Abdulridha, O. Batuman, and Y. Ampatzidis, ‘‘UAV-based remote [247] G. Polder, N. V. D. Westeringh, J. Kool, H. A. Khan, G. Kootstra, and
sensing technique to detect citrus canker disease utilizing hyperspectral A. Nieuwenhuizen, ‘‘Automatic detection of tulip breaking virus (TBV)
imaging and machine learning,’’ Remote Sens., vol. 11, no. 11, p. 1373, using a deep convolutional neural network,’’ IFAC-PapersOnLine, vol. 52,
Jun. 2019. no. 30, pp. 12–17, 2019.
[229] C. Nguyen, V. Sagan, M. Maimaitiyiming, M. Maimaitijiang, S. Bhadra, [248] E. C. Oerke, A. K. Mahlein, and U. Steiner, ‘‘Proximal sensing of plant
and M. T. Kwasniewski, ‘‘Early detection of plant viral disease using diseases,’’ in Detection and Diagnostics of Plant Pathogens, vol. 5,
hyperspectral imaging and deep learning,’’ Sensors, vol. 21, no. 3, p. 742, M. Gullino and P. Bonants, Eds. Dordrecht, The Netherlands: Springer,
Jan. 2021. 2014.
[230] L. Feng, B. Wu, Y. He, and C. Zhang, ‘‘Hyperspectral imaging combined [249] S. G. Caro. (Jan. 21, 2014). Infection and Spread of Peronospora
with deep transfer learning for rice disease detection,’’ Frontiers Plant Sparsa on Rosa SP. (Berk). Accessed: Jan. 8, 2024. [Online]. Available:
Sci., vol. 12, Sep. 2021, Art. no. 693521. https://bonndoc.ulb.uni-bonn.de/xmlui/handle/20.500.11811/5823
[231] L. Wan, H. Li, C. Li, A. Wang, Y. Yang, and P. Wang, ‘‘Hyperspectral [250] E. Bauriegel and W. Herppich, ‘‘Hyperspectral and chlorophyll
sensing of plant diseases: Principle and methods,’’ Agronomy, vol. 12, fluorescence imaging for early detection of plant diseases,
no. 6, p. 1451, Jun. 2022, doi: 10.3390/agronomy12061451. with special reference to fusarium spec. Infections on wheat,’’
[232] V. G. Bhujade and V. Sambhe, ‘‘Role of digital, hyper spectral, and Agriculture, vol. 4, no. 1, pp. 32–57, Mar. 2014, doi: 10.3390/
SAR images in detection of plant disease with deep learning network,’’ agriculture4010032.
Multimedia Tools Appl., vol. 81, no. 23, pp. 33645–33670, Apr. 2022, doi: [251] C. Rousseau, E. Belin, E. Bove, D. Rousseau, F. Fabre, R. Berruyer,
10.1007/s11042-022-13055-z. J. Guillaumès, C. Manceau, M.-A. Jacques, and T. Boureau, ‘‘High
[233] R. Cui, J. Li, Y. Wang, S. Fang, K. Yu, and Y. Zhao, ‘‘Hyperspec- throughput quantitative phenotyping of plant resistance using chloro-
tral imaging coupled with dual-channel convolutional neural network phyll fluorescence image analysis,’’ Plant Methods, vol. 9, no. 1, p. 17,
for early detection of apple valsa canker,’’ Comput. Electron. Agricult., Dec. 2013, doi: 10.1186/1746-4811-9-17.
vol. 202, Nov. 2022, Art. no. 107411, doi: 10.1016/j.compag.2022. [252] Y. Li and X. Chao, ‘‘Semi-supervised few-shot learning approach for
107411. plant diseases recognition,’’ Plant Methods, vol. 17, no. 1, pp. 1–10,
[234] Y. Wu, Y. Cao, and Z. Zhai, ‘‘Early detection of bacterial blight in Jun. 2021, doi: 10.1186/s13007-021-00770-1.
hyperspectral images based on random forest and adaptive coherence [253] M. Buda, A. Maki, and M. A. Mazurowski, ‘‘A systematic study of
estimator,’’ Sustainability, vol. 14, no. 20, p. 13168, Oct. 2022, doi: the class imbalance problem in convolutional neural networks,’’ Neural
10.3390/su142013168. Netw., vol. 106, pp. 249–259, Oct. 2018.
[235] J. Ugarte Fajardo, M. Maridueña-Zavala, J. Cevallos-Cevallos, and [254] J. M. Johnson and T. M. Khoshgoftaar, ‘‘Survey on deep learn-
D. O. Donoso, ‘‘Effective methods based on distinct learning principles ing with class imbalance,’’ J. Big Data, vol. 6, no. 1, pp. 1–54,
for the analysis of hyperspectral images to detect black Sigatoka disease,’’ Dec. 2019.
Plants, vol. 11, no. 19, p. 2581, Sep. 2022, doi: 10.3390/plants11192581. [255] A. Medela, A. Picon, C. L. Saratxaga, O. Belar, V. Cabezon, R. Cicchi,
[236] L. Z. Yong, S. Khairunniza-Bejo, M. Jahari, and F. M. Muharam, R. Bilbao, and B. Glover, ‘‘Few shot learning in histopathological
‘‘Automatic disease detection of basal stem rot using deep learn- images:Reducing the need of labeled data on biological datasets,’’ in Proc.
ing and hyperspectral imaging,’’ Agriculture, vol. 13, no. 1, p. 69, IEEE 16th Int. Symp. Biomed. Imag. (ISBI ), Venice, Italy, Apr. 2019,
Dec. 2022. pp. 1860–1864.
[237] G. Wu, Y. Fang, Q. Jiang, M. Cui, N. Li, Y. Ou, Z. Diao, and B. Zhang, [256] S.-N. Ren, Y. Sun, H.-Y. Zhang, and L.-X. Guo, ‘‘Plant disease identifi-
‘‘Early identification of strawberry leaves disease utilizing hyperspec- cation for small sample based on one-shot learning,’’ Jiangsu J. Agricult.
tral imaging combing with spectral features, multiple vegetation indices Sci., vol. 35, no. 5, pp. 1061–1067, May 2019.
and textural features,’’ Comput. Electron. Agricult., vol. 204, Jan. 2023, [257] D. Argüeso, A. Picon, U. Irusta, A. Medela, M. G. San-Emeterio,
Art. no. 107553. A. Bereciartua, and A. Alvarez-Gila, ‘‘Few-shot learning approach for
[238] A.-K. Mahlein, ‘‘Plant disease detection by imaging sensors—Parallels plant disease classification using images taken in the field,’’ Comput.
and specific demands for precision agriculture and plant phenotyping,’’ Electron. Agricult., vol. 175, Aug. 2020, Art. no. 105542.
Plant Disease, vol. 100, no. 2, pp. 241–251, Feb. 2016, doi: 10.1094/pdis- [258] D. Das and C. S. G. Lee, ‘‘A two-stage approach to few-shot learn-
03-15-0340-fe. ing for image recognition,’’ IEEE Trans. Image Process., vol. 29,
[239] A.-K. Mahlein, M. T. Kuska, J. Behmann, G. Polder, and A. Walter, pp. 3336–3350, 2020.
‘‘Hyperspectral sensors and imaging technologies in phytopathology: [259] F. Zhong, Z. Chen, Y. Zhang, and F. Xia, ‘‘Zero- and few-shot learning
State of the art,’’ Annu. Rev. Phytopathol., vol. 56, no. 1, pp. 535–558, for diseases recognition of citrus aurantium L. Using conditional adver-
Aug. 2018, doi: 10.1146/annurev-phyto-080417-050100. sarial autoencoders,’’ Comput. Electron. Agricult., vol. 179, Dec. 2020,
[240] S. K. von Bueren, A. Burkart, A. Hueni, U. Rascher, M. P. Tuohy, and Art. no. 105828.
I. J. Yule, ‘‘Deploying four optical UAV-based sensors over grassland: [260] H.-Y. Wu, ‘‘Identification of tea leaf’s diseases in natural scene images
Challenges and limitations,’’ Biogeosciences, vol. 12, no. 1, pp. 163–175, based on low shot learning,’’ M.S. thesis, Dept. Inf. Eng., Anhui Univ.,
Jan. 2015, doi: 10.5194/bg-12-163-2015. Hefei, China, 2020.
[241] W. Lv and X. Wang, ‘‘Overview of hyperspectral image classification,’’ [261] Y. Li and J. Yang, ‘‘Meta-learning baselines and database for few-
J. Sensors, vol. 2020, pp. 1–13, Jul. 2020, doi: 10.1155/2020/4817234. shot classification in agriculture,’’ Comput. Electron. Agricult., vol. 182,
[242] J.-H. Lenthe, E.-C. Oerke, and H.-W. Dehne, ‘‘Digital infrared thermog- Mar. 2021, Art. no. 106055, doi: 10.1016/j.compag.2021.106055.
raphy for monitoring canopy health of wheat,’’ Precis. Agricult., vol. 8, [262] X. Liang, ‘‘Few-shot cotton leaf spots disease classification based on
nos. 1–2, pp. 15–26, Mar. 2007, doi: 10.1007/s11119-006-9025-6. metric learning,’’ Plant Methods, vol. 17, no. 1, pp. 1–11, Nov. 2021, doi:
[243] E. Bauriegel, H. Brabandt, U. Gärber, and W. B. Herppich, ‘‘Chlorophyll 10.1186/s13007-021-00813-7.
fluorescence imaging to facilitate breeding of bremia lactucae-resistant [263] S. Wang, Y. Han, J. Chen, X. He, Z. Zhang, X. Liu, and K. Zhang,
lettuce cultivars,’’ Comput. Electron. Agricult., vol. 105, pp. 74–82, ‘‘Weed density extraction based on few-shot learning through UAV
Jul. 2014, doi: 10.1016/j.compag.2014.04.010. remote sensing RGB and multispectral images in ecological irrigation
[244] M. Neumann, L. Hallau, B. Klatt, K. Kersting, and C. Bauckhage, area,’’ Frontiers Plant Sci., vol. 12, Mar. 2022, Art. no. 735230, doi:
‘‘Erosion band features for cell phone image based plant disease classifi- 10.3389/fpls.2021.735230.
cation,’’ in Proc. 22nd Int. Conf. Pattern Recognit., Stockholm, Sweden, [264] I. Egusquiza, A. Picon, U. Irusta, A. Bereciartua-Perez, T. Eggers,
Aug. 2014, pp. 3315–3320, doi: 10.1109/ICPR.2014.571. C. Klukas, E. Aramendi, and R. Navarra-Mestre, ‘‘Analysis of few-
[245] S. Bergstrasser, D. Fanourakis, S. Schmittgen, M. P. Cendrero-Mateo, shot techniques for fungal plant disease classification and evalua-
M. Jansen, H. Scharr, and U. Rascher, ‘‘HyperART: Non-invasive tion of clustering capabilities over real datasets,’’ Frontiers Plant
quantification of leaf traits using hyperspectral absorption-reflectance- Sci., vol. 13, Mar. 2022, Art. no. 813237, doi: 10.3389/fpls.2022.
transmittance imaging,’’ Plant Methods, vol. 11, pp. 1–17, Dec. 2015. 813237.
VOLUME 12, 2024 37467

[265] S. Garg and P. Singh, ‘‘An aggregated loss function based lightweight few [287] L. C. Clark and C. Lyons, ‘‘Electrode systems for continuous monitoring
shot model for plant leaf disease classification,’’ Multimedia Tools Appl., in cardiovascular surgery,’’ Ann. New York Acad. Sci., vol. 102, no. 1,
vol. 82, no. 15, pp. 23797–23815, Feb. 2023, doi: 10.1007/s11042-023- pp. 29–45, Oct. 1962.
14372-7. [288] F. Sanger, S. Nicklen, and A. R. Coulson, ‘‘DNA sequencing with chain-
[266] X. Liu, Z. Ji, Y. Pang, and Z. Han, ‘‘Self-taught cross-domain terminating inhibitors,’’ Proc. Nat. Acad. Sci. USA, vol. 74, no. 12,
few-shot learning with weakly supervised object localization and pp. 5463–5467, Dec. 1977.
task-decomposition,’’ Knowl.-Based Syst., vol. 265, Apr. 2023, [289] L. Liu, Y. Li, S. Li, N. Hu, Y. He, R. Pong, D. Lin, L. Lu,
Art. no. 110358, doi: 10.1016/j.knosys.2023.110358. and M. Law, ‘‘Comparison of next-generation sequencing systems,’’
[267] P. Li, F. Liu, L. Jiao, S. Li, L. Li, X. Liu, and X. Huang, ‘‘Knowledge trans- J. Biomed. Biotechnol., vol. 2012, pp. 1–11, Oct. 2012.
duction for cross-domain few-shot learning,’’ Pattern Recognit., vol. 141, [290] A. M. Kanzi, J. E. San, B. Chimukangara, E. Wilkinson, M. Fish,
Sep. 2023, Art. no. 109652, doi: 10.1016/j.patcog.2023.109652. V. Ramsuran, and T. de Oliveira, ‘‘Next generation sequencing and
[268] M. A. Chandra and S. S. Bedi, ‘‘Survey on SVM and their application bioinformatics analysis of family genetic inheritance,’’ Frontiers Genet.,
in image classification,’’ Int. J. Inf. Technol., vol. 13, no. 5, pp. 1–11, vol. 11, Oct. 2020, Art. no. 544162.
Oct. 2021. [291] J. Qin, R. Li, J. Raes, M. Arumugam, K. S. Burgdorf, C. Manichanh,
[269] S. Vandenhende, S. Georgoulis, W. Van Gansbeke, M. Proesmans, T. Nielsen, N. Pons, F. Levenez, T. Yamada, and D. R. Mende, ‘‘A human
D. Dai, and L. Van Gool, ‘‘Multi-task learning for dense prediction gut microbial gene catalogue established by metagenomic sequencing,’’
tasks: A survey,’’ IEEE Trans. Pattern Anal. Mach. Intell., vol. 44, Nature, vol. 464, no. 7285, pp. 59–65, 2010.
no. 7, pp. 3614–3633, Jul. 2022. [Online]. Available: https://homes. [292] S. Darmanis, R. Y. Nong, J. Vänelid, A. Siegbahn, O. Ericsson,
esat.kuleuven.be/~konijn/publications/2020/SVDH.pdf S. Fredriksson, C. Bäcklin, M. Gut, S. Heath, I. G. Gut, L. Wallentin,
[270] S. Bozinovski, ‘‘Reminder of the first paper on transfer learning in neural M. G. Gustafsson, M. Kamali-Moghaddam, and U. Landegren, ‘‘Pro-
networks, 1976,’’ Informatica, vol. 44, no. 3, pp. 291–302, Sep. 2020. teinSeq: High-performance proteomic analyses by proximity ligation
[271] T. Uehara-Ichiki, T. Shiba, K. Matsukura, T. Ueno, M. Hirae, and and next generation sequencing,’’ PLoS ONE, vol. 6, no. 9, Sep. 2011,
T. Sasaya, ‘‘Detection and diagnosis of rice-infecting viruses,’’ Frontiers Art. no. e25583.
Microbiol., vol. 4, p. 289, Oct. 2013. [293] T. M. Walker, C. L. Ip, R. H. Harrell, J. T. Evans, G. Kapatai,
[272] T. Stackhouse, A. D. Martinez-Espinoza, and M. E. Ali, ‘‘Turfgrass dis- M. J. Dedicoat, D. W. Eyre, D. J. Wilson, P. M. Hawkey, D. W. Crook,
ease diagnosis: Past, present, and future,’’ Plants, vol. 9, no. 11, p. 1544, J. Parkhill, D. Harris, A. S. Walker, R. Bowden, P. Monk, E. G. Smith,
Nov. 2020. and T. E. Peto, ‘‘Whole-genome sequencing to delineate mycobacterium
[273] S. Sakamoto, W. Putalun, S. Vimolmangkang, W. Phoolcharoen, tuberculosis outbreaks: A retrospective observational study,’’ Lancet
Y. Shoyama, H. Tanaka, and S. Morimoto, ‘‘Enzyme-linked immunosor- Infectious Diseases, vol. 13, no. 2, pp. 137–146, Feb. 2013.
bent assay for the quantitative/qualitative analysis of plant secondary [294] G. A. Díaz-Cruz, C. M. Smith, K. F. Wiebe, S. M. Villanueva,
metabolites,’’ J. Natural Medicines, vol. 72, no. 1, pp. 32–42, Jan. 2018. A. R. Klonowski, and B. J. Cassone, ‘‘Applications of next-generation
[274] H. A. McCartney, S. J. Foster, B. A. Fraaije, and E. Ward, ‘‘Molecular sequencing for large-scale pathogen diagnoses in soybean,’’ Plant Dis-
diagnostics for fungal plant pathogens,’’ Pest Manag. Sci., vol. 59, no. 2, ease, vol. 103, no. 6, pp. 1075–1083, Jun. 2019.
pp. 129–142, Feb. 2003. [295] B. E. Slatko, A. F. Gardner, and F. M. Ausubel, ‘‘Overview of next-
[275] S. Loreti, ‘‘The diagnosis of plant pathogenic bacteria a state of art,’’ generation sequencing technologies,’’ Curr. Protoc. Mol. Biol., vol. 122,
Frontiers Bioscience, vol. 10, no. 3, pp. 449–460, 2018. no. 1, p. e59, 2018.
[276] K. B. Mullis and F. A. Faloona, ‘‘Specific synthesis of DNA in vitro [296] T. H. Meen, ‘‘IoT, communication, and engineering,’’ in Proc. IEEE
via a polymerase-catalyzed chain reaction,’’ in Methods in Enzymology, Eurasia Conf. IoT, Commun., Eng., vol. 3, Yunlin, Taiwan, Oct. 2019.
vol. 155, San Diego, CA, USA: Elsevier, 1987, pp. 335–350. [297] S. M. M. Hossain, K. Deb, P. K. Dhar, and T. Koshiba, ‘‘Plant leaf dis-
[277] P. Bastien, G. W. Procop, and U. Reischl, ‘‘Quantitative real-time PCR is ease recognition using depth-wise separable convolution-based models,’’
not more sensitive than ‘conventional’ PCR,’’ J. Clin. Microbiol., vol. 46, Symmetry, vol. 13, no. 3, pp. 511–529, Mar. 2021.
no. 6, pp. 1897–1900, Jun. 2008.
[278] Lsuhsc.edu. Accessed: Aug. 10, 2023. [Online]. Available: https://
nursing.lsuhsc.edu/JBI/docs/JBIBooks/Diagnostic%20Accuracy.pdf
[279] J. M. Crosslin, G. J. Vandemark, and J. E. Munyaneza, ‘‘Development of
a real-time, quantitative PCR for detection of the Columbia basin potato ANUJA BHARGAVA received the Ph.D. degree
purple top phytoplasma in plants and beet leafhoppers,’’ Plant Disease, in electronics and communication from GLA Uni-
vol. 90, no. 5, pp. 663–667, May 2006. versity, Mathura, Uttar Pradesh, in 2020. She is
[280] C. J. Smith and A. M. Osborn, ‘‘Advantages and limitations of quantitative currently an Assistant Professor with the Depart-
PCR (Q-PCR)-based approaches in microbial ecology,’’ FEMS Micro- ment of Electronics and Communication, Institute
biol. Ecol., vol. 67, no. 1, pp. 6–20, Jan. 2009. of Engineering and Technology, GLA University.
[281] T. Notomi, ‘‘Loop-mediated isothermal amplification of DNA,’’ Nucleic She has more than 14 years of teaching experience.
Acids Res., vol. 28, no. 12, p. 63, Jun. 2000. Her current research interests include image pro-
[282] E. S. M. Tuppurainen, E. H. Venter, J. L. Shisler, G. Gari, cessing, machine learning, and computer vision.
G. A. Mekonnen, N. Juleff, N. A. Lyons, K. De Clercq, C. Upton,
She has more than 25 publications in international
T. R. Bowden, S. Babiuk, and L. A. Babiuk, ‘‘Review: Capripoxvirus
journals and conference of repute.
diseases: Current status and opportunities for control,’’ Transboundary
Emerg. Diseases, vol. 64, no. 3, pp. 729–745, Jun. 2017.
[283] P. Hardinge and J. A. H. Murray, ‘‘Reduced false positives and improved
reporting of loop-mediated isothermal amplification using quenched flu-
orescent primers,’’ Sci. Rep., vol. 9, no. 1, pp. 1–13, May 2019.
AASHEESH SHUKLA received the Ph.D. degree
[284] S. Waliullah, K.-S. Ling, E. J. Cieniewicz, J. E. Oliver, P. Ji, and M. E. Ali,
in electronics and communication from GLA Uni-
‘‘Development of loop-mediated isothermal amplification assay for rapid
versity, Mathura, Uttar Pradesh, in 2017. He is
detection of cucurbit leaf crumple virus,’’ Int. J. Mol. Sci., vol. 21, no. 5,
p. 1756, Mar. 2020. currently a Professor with the Department of
[285] A. Rani, N. Donovan, and N. Mantri, ‘‘Review: The future of plant Electronics and Communication, GLA University
pathogen diagnostics in a nursery production system,’’ Biosensors Bio- of Technology and Management. He has more
electron., vol. 145, Dec. 2019, Art. no. 111631. than 15 years of teaching experience. His current
[286] M. M. F. Azizi, N. H. Mardhiah, and H. Y. Lau, ‘‘Current and emerg- research interests include wireless communica-
ing molecular technologies for the diagnosis of plant diseases—An tion, machine learning, and image processing.
overview,’’ J. Experim. Biol. Agricult. Sci., vol. 10, no. 2, pp. 294–305, He has more than 30 publications in international
Apr. 2022. journals and conference of repute.
37468 VOLUME 12, 2024

OM PRAKASH GOSWAMI holds the Ph.D. PEERAPONG UTHANSAKUL (Member, IEEE)

degree in signal processing from the Netaji Sub- received the B.Eng. and M.Eng. degrees in elec-
has Institute of Technology, University of Delhi, trical engineering from Chulalongkorn University,
India. He is working as an Assistant Professor Thailand, in 1996 and 1998, respectively, and the
with Somaiya Vidyavihar University, Mumbai. His Ph.D. degree in information technology and elec-
research interests include digital signal processing trical engineering from The University of Queens-
and fractional-filters. He has co-authored several land, Australia, in 2007. From 1998 to 2001,
reputed research papers. he was a Telecommunication Engineer in one of
the leading telecommunication company of Thai-
land. He has been an Associate Professor and the
Director of the Research and Development Institute, Suranaree University of
Technology, Thailand. He has got more than 300 research publications and
he is the author/coauthor of various books related to MIMO technologies.
He has won various national awards from the government of Thailand due
to his contributions and motivation in the field of Science and Technology.
His research interests include green communications, wave propagation
modeling, wireless communications, MIMO, massive MIMO, brain wave
engineering, mobile networks, and the quality of experience (QoE).
MOHAMMED H. ALSHARIF received the
B.Eng. degree in electrical engineering (wireless
communication and networking) from the Islamic
University of Gaza, Palestine, in 2008, and the
M.Sc.Eng. and Ph.D. degrees in electrical engi- MONTHIPPA UTHANSAKUL (Member, IEEE)
neering (wireless communication and network- received the B.E. degree in telecommunication
ing) from the National University of Malaysia, engineering from the Suranaree University of
Malaysia, in 2012 and 2015, respectively. In Technology, Thailand, in 1997, the M.E. degree
2016, he joined the Faculty of Sejong University, in electrical engineering from Chulalongkorn Uni-
South Korea, as an Assistant Professor with the versity, Thailand, in 1999, and the Ph.D. degree
Department of Electrical Engineering, where he has since been promoted in information technology and electrical engi-
to an Associate Professor. His current research interests include wireless neering from The University of Queensland,
communications and networks, including cutting-edge areas, such as wire- Australia, in 2007. She is currently an Associate
less communications, network information theory, the Internet of Things Professor with the Telecommunication School,
(IoT), green communication, energy-efficient wireless transmission tech- Suranaree University of Technology. Her research interests include wide-
niques, wireless power transfer, and wireless energy harvesting. His research band/narrowband smart antennas, automatic switch beam antenna, DOA
achievements are reflected in his extensive publication record in top-tier jour- finder, microwave components, the application of smart antenna, and
nals in electrical and electronics/communications engineering. His expertise advance wireless communications. She received the Second Prize of Young
and contributions have also been recognized by leading publishers, such as Scientist Award from 16th International Conference on Microwaves, Radar,
IEEE, Elsevier, Springer, and MDPI, who has invited him to serve as a guest and Wireless Communications, Poland, in 2006.
editor for many special issues.
VOLUME 12, 2024 37469

Plant Leaf Disease Detection Classification and Diagnosis Using Computer Vision and Artificial Intelligence A Review

Uploaded by

Copyright:

Available Formats

Plant Leaf Disease Detection Classification and Diagnosis Using Computer Vision and Artificial Intelligence A Review

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Plant Leaf Disease Detection Classification and Diagnosis Using Computer Vision and Artificial Intelligence A Review

Uploaded by

Copyright:

Available Formats

Received 4 February 2024, accepted 26 February 2024, date of publication 4 March 2024, date of current version 14 March 2024.

Digital Object Identifier 10.1109/ACCESS.2024.3373001

Plant Leaf Disease Detection, Classification, and

Corresponding authors: Mohammed H. Alsharif (malsharif@sejong.ac.kr) and Peerapong Uthansakul (uthansakul@sut.ac.th)

I. INTRODUCTION than 85% of people confide in agriculture in the world [2].

detection, management, and prevention of disease in plants

37444 VOLUME 12, 2024

FIGURE 3. Classification of plant disease in distinct categories [16].

TABLE 1. Distinct disease in different plants.

VOLUME 12, 2024 37445

TABLE 2. Distinct plants, their disease and responsible pathogen [16].

37446 VOLUME 12, 2024

VOLUME 12, 2024 37447

before classification. The features are advantageous to ana-

37448 VOLUME 12, 2024

B. DEEP LEARNING USING IMAGE PROCESSING

VOLUME 12, 2024 37449

37450 VOLUME 12, 2024

FIGURE 7. Hierarchy of DL techniques.

VOLUME 12, 2024 37451

TABLE 7. Details of classification technique used by various researchers.

37452 VOLUME 12, 2024

TABLE 7. (Continued.) Details of classification technique used by various researchers.

VOLUME 12, 2024 37453

TABLE 7. (Continued.) Details of classification technique used by various researchers.

FIGURE 8. Elemental steps of deep learning.

Goodfellow et al. 2014 [167] propose a generating model

37454 VOLUME 12, 2024

TABLE 11. Details of DL technique used by various researchers.

FIGURE 9. Generate adversarial network architecture.

TABLE 10. Various literature regarding data augmentation.

C. HYPERSPECTRAL IMAGING USING IMAGE

VOLUME 12, 2024 37455

An extensive number of modern developments in different

37456 VOLUME 12, 2024

VOLUME 12, 2024 37457

FIGURE 11. Overview of sensor technologies [238].

FIGURE 12. Approach to imitate multiple sources.

37458 VOLUME 12, 2024

FIGURE 13. Illustrative real-time PCR schematic [286].

FIGURE 14. Amplification of LAMP [283].

FIGURE 15. Biosensors & its process schematic illustration [286].

VOLUME 12, 2024 37459

37460 VOLUME 12, 2024

VOLUME 12, 2024 37461

37462 VOLUME 12, 2024

VOLUME 12, 2024 37463

37464 VOLUME 12, 2024

VOLUME 12, 2024 37465

37466 VOLUME 12, 2024

VOLUME 12, 2024 37467

37468 VOLUME 12, 2024

OM PRAKASH GOSWAMI holds the Ph.D. PEERAPONG UTHANSAKUL (Member, IEEE)

VOLUME 12, 2024 37469

You might also like