Next Article in Journal
Technology Transfer Assessment in Regional Business Contexts
Previous Article in Journal
The Restorative and Contingent Value of Biophilic Indoor Environments in Healthcare Settings
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Sustainability
Volume 15
Issue 15
10.3390/su151511681
sustainability-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Review

Transformative Role of Artificial Intelligence in Advancing Sustainable Tomato (Solanum lycopersicum) Disease Management for Global Food Security: A Comprehensive Review

by
Bharathwaaj Sundararaman
1,†,
Siddhant Jagdev
2,† and
Narendra Khatri
3,*
1
Department of Electrical and Electronics Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal 576104, India
2
Department of Instrumentation and Control Engineering, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal 576104, India
3
Department of Mechatronics, Manipal Institute of Technology, Manipal Academy of Higher Education, Manipal 576104, India
*
Author to whom correspondence should be addressed.
†
These authors contributed equally to this work and should be considered co-first authors.
Sustainability 2023, 15(15), 11681; https://doi.org/10.3390/su151511681
Submission received: 13 June 2023 / Revised: 17 July 2023 / Accepted: 19 July 2023 / Published: 28 July 2023
(This article belongs to the Section Sustainable Food)
Browse Figures
Versions Notes

Abstract

:
The growing global population and accompanying increase in food demand has put pressure on agriculture to produce higher yields in the face of numerous challenges, including plant diseases. Tomato is a widely cultivated and essential food crop that is particularly susceptible to disease, resulting in significant economic losses and hindrances to food security. Recently, Artificial Intelligence (AI) has emerged as a promising tool for detecting and classifying tomato leaf diseases with exceptional accuracy and efficiency, empowering farmers to take proactive measures to prevent crop damage and production loss. AI algorithms are capable of processing vast amounts of data objectively and without human bias, making them a potent tool for detecting even subtle variations in plant diseases that traditional techniques might miss. This paper provides a comprehensive overview of the most recent advancements in tomato leaf disease classification using Machine Learning (ML) and Deep Learning (DL) techniques, with an emphasis on how these approaches can enhance the accuracy and effectiveness of disease classification. Several ML and DL models, including convolutional neural networks (CNN), are evaluated for tomato leaf disease classification. This review paper highlights the various features and techniques used in data acquisition as well as evaluation metrics employed to assess the performance of these models. Moreover, this paper emphasizes how AI techniques can address the limitations of traditional techniques in tomato leaf disease classification, leading to improved crop yields and more efficient management techniques, ultimately contributing to global food security. This review paper concludes by outlining the limitations of recent research and proposing new research directions in the field of AI-assisted tomato leaf disease classification. These insights will be of significant value to researchers and professionals interested in utilizing ML and DL techniques for tomato leaf disease classification and ultimately contribute to sustainable food production (SDG-3).

1. Introduction

The escalating imperative for sustainable food security has spurred the integration of artificial intelligence (AI) into agriculture for efficacious problem-solving. In recent times, an unprecedented upsurge has been witnessed in the utilization of AI algorithms for discerning and classifying plant diseases, with a particular focus on Solanum lycopersicum, commonly known as tomatoes [1]. This remarkable progress can be ascribed to the exponential proliferation of data availability, concomitant with groundbreaking advancements in AI algorithms and computer vision, culminating in the development of highly refined algorithms that exhibit unparalleled accuracy for precisely identifying and categorizing diseases in tomato plants [2]. The application of AI in plant disease detection and classification holds prodigious potential in augmenting crop management practices, mitigating crop losses, and ameliorating food security (Sustainable Development Goal-3). Thus, the indispensability of AI in agriculture is unequivocally evident in fostering sustainable food production and fulfilling the burgeoning global demand for food.
Plant disease classification has been a prominent focus of research in machine learning (ML), with techniques such as random forest (RF), decision trees, and support vector machines (SVM) commonly employed [3]. However, these methods face challenges in effectively recognizing intricate image patterns while managing large datasets. To overcome these limitations, convolutional neural networks (CNNs), a subset of deep learning (DL) techniques, have been developed and have demonstrated significant advancements in plant disease classification [4].
In a study by Mohanty et al. (2016), the accuracy of deep learning techniques for plant disease classification was evaluated using the PlantVillage dataset, yielding highly promising results. Across all experimental setups, accuracy ranged from 85.53% to 99.34%, with negligible differences between training loss and validation loss, indicating minimal overfitting. Notably, the GoogLeNet architecture consistently outperformed the AlexNet architecture, and transfer learning showed superior results depending on the training strategy. Remarkably, experiments conducted with colored images exhibited the best performance. These findings underscore the tremendous potential of deep learning techniques, particularly those employing the GoogLeNet architecture and transfer learning, for precise and reliable plant disease classification [5,6,7].
In 2017, Fuentes et al. introduced a robust deep learning (DL)-based detector for real-time identification of tomato diseases and pests [8]. This approach offers an adaptable and effective solution for identifying the class of disease and pests and locating them in images. The detector can analyze 6.25 images per second, making it a fast and efficient method for disease and pest detection [9,10].
Li et al. (2023) introduced a tomato leaf disease identification method called LMBRNet. LMBRNet utilizes Complementary Grouped Dilated Residual feature extraction blocks and employs four branches with convolution kernels of different sizes to capture diverse characteristics of tomato leaf diseases [11]. Sunil C.K. et al. (2023) developed a classification model for tomato plant disease detection that combines MFFN and ACSPAM. An accuracy rate of 99.88% was attained by the model during training, validation, and testing. It also included recommendations for pesticides based on the illness categorization, which is a unique addition [12].
Sanida et al. (2023) developed a unique network for detecting tomato leaf illnesses by merging VGG16’s first ten convolution layers and two inception blocks. To mitigate overfitting, transfer learning is applied, and a two-phase progressive model training approach is implemented. When compared to previous methodologies, the model has a superior accuracy rate of 99.23% [13].
The system analyzes real-time images using GPUs instead of collecting physical samples. The dataset incorporates environmental complexity variables. The DL-based detector accurately identifies nine disease and pest categories, outperforming other architectures. Data annotation and augmentation techniques enhance performance [14].
The progress made in the existing literature on the classification of tomato leaf diseases using artificial intelligence (AI) techniques has been significant, largely due to the integration of AI algorithms and advancements in computer vision. Despite the growing body of research in this area, there remains a need to identify the gaps, patterns, and challenges in recognizing and classifying tomato leaf diseases. To address these gaps, this review aims to comprehensively evaluate the literature and answer six research questions (RQs). These RQs encompass analyzing the contributions of current studies, understanding the diseases affecting tomato crops, categorizing the datasets utilized, identifying the machine learning (ML) and deep learning (DL) algorithms employed, determining the features used for recognition and classification, and deducing the challenges and trends in tomato plant disease detection. By addressing these RQs, this review will contribute to the advancement of AI for achieving Sustainable Development Goal 3 (SDG-3) related to good health and well-being [15]. Table 1 outlines the research questions.
The research questions outlined in Table 1 will facilitate the development of more effective ML/DL algorithms for the automatic classification of tomato plant diseases. By addressing these research questions, this review will help to enhance agricultural productivity.
The remaining sections of this paper are organized as follows: Section 2 presents the theoretical background. Section 3 discusses the review methodology used for the research. Section 4 provides a description of nine common diseases that affect tomato plants. Section 5 discusses various datasets available for tomato disease classification. Section 6 summarizes different algorithms used for plant disease detection and classification. Section 7 presents an overview of the various features used to identify different diseases. Section 8 presents the discussion on challenges and trends identified in the review process regarding tomato plant disease detection and classification. Finally, Section 9 provides concluding remarks on the review.

2. Theoretical Background

The existing literature on the classification of tomato leaf diseases using artificial intelligence (AI) techniques has witnessed significant advancements, particularly with the application of machine learning (ML) and deep learning (DL) algorithms. However, there remains a need to comprehensively evaluate the contributions, datasets, algorithms, and features employed in these studies. Additionally, the current literature lacks a thorough analysis of the challenges and trends in tomato plant disease detection. By addressing these gaps, this review aims to provide a comprehensive understanding of the current state-of-the-art techniques and identify areas for further research. This research will contribute to the development of more effective ML/DL algorithms for automatic and precise classification of tomato plant diseases, thereby enhancing agricultural productivity and promoting sustainable food security.
In 2018, Ferentinos trained DL CNN models to diagnose plant diseases using basic leaf photos of healthy and ill plants. The training data included an open collection of 87,848 photos of 25 plants in 58 classes of plant-disease pairings, including healthy plants. The VGG CNN model showed the best performance, with 99.53% accuracy in detecting both diseased and healthy plants. The study highlighted the importance of field-taken photos in the training data and emphasized the need to increase their proportion when constructing such models [16,17].
In 2018, Fuentes et al. created a real-time tomato plant disease and pest detection system using object-specific neural networks. The system employed bounding boxes to precisely recognize images from multiple field cameras. The model exhibited real-time detection of nearly 96% of diseases and pests [18].
In 2019, Pravin Kumar et al. developed a multi-kernel with parallel DL (MK-PDL) classifier to detect diseased leaves and compared it to other classifiers. The MK-PDL classifier achieved superior performance with high accuracy, specificity, sensitivity, F-measure, recall, and precision rates, indicating its potential to accurately identify diseased leaves in real-world agricultural environments [19].
In 2019, KC et al. investigated the performance of various model architectures in identifying plant diseases and found that MobileNet was the most effective, achieving a high success rate with significantly fewer parameters than VGG [20]. The researchers also found that a reduced version of MobileNet achieved similar accuracy levels with fewer parameters, highlighting the negligible contribution of the network’s final repeating layers [21].
In 2019, G. et al. proposed a Deep CNN model for the accurate classification of 38 distinct types of healthy and unhealthy plants. The model was trained and tested on a large dataset and demonstrated high accuracy, with an average accuracy of 96.46% on the testing set and high accuracy rates for each class [22]. Multiple feature extraction methods were utilized to detect four tomato leaf diseases with high accuracy using decision tree and random forest classifiers, with the random forest classifier showing better accuracy [23].
In 2020, Chen et al. employed ABCK-BWTR and B-ARNet models to detect tomato leaf disease with an accuracy of 89% [24]. R. et al. proposed a residual network with an attention mechanism for deep learning, which achieved a validation set accuracy of 98% for tomato leaf disease identification [25].
In 2020, Li et al. presented a SE-Inception model for identifying solanaceous diseases using mobile devices. The model incorporated a multi-scale feature extraction module and a SE module for effective channel information exploitation, resulting in high recognition accuracies on their dataset and the PlantVillage public dataset, with small model size [26].
Tan et al. (2021) conducted a comparative study of machine learning (ML) and deep learning (DL) methods for plant disease identification using the PlantVillage dataset [27]. They extracted texture features and color characteristics from the images and evaluated their performance on RF, kNN, SVM, EfficientNet-b0, VGG16, ResNet34, AlexNet, and MobileNetV2. The COLOR + GLCM approach achieved the best results, and the ResNet34 network outperformed other ML/DL algorithms, achieving a 99.7% F1 score, 99.6% precision, 99.7% recall, and 99.7% accuracy [6,16,28,29,30].
Peker (2021) proposed a novel deep learning technique called the Multi-channel Capsule Network Ensemble for plant disease detection. This approach combines ensemble learning with a five-channel capsule network, which draws information from five distinct data sources. The ensembled capsule network model outperformed all other models that have been previously reported in the literature, demonstrating its superior performance [31].
Wang et al. (2022) developed a dual-stream hierarchical bilinear pooling model for multi-task plant and disease classification. The model leverages multi-layer information to accurately classify both plants and diseases, achieving plant classification accuracy of 84.71% and disease classification accuracy of 75.06% on a real-world dataset. The researchers attribute the model’s improved representation ability to enhanced information interaction between network layers and learning of distinguishing features [32].
Yu et al. (2023) proposed an advanced plant disease diagnosis deep learning network called ICVT (Inception Convolutional Vision Transformer), which captures both local and high-level plant disease detection information. The network achieved state-of-the-art performance with an average accuracy of 86.89%, 99.94%, 99.22%, and 77.54% on AI2018, PlantVillage, ibean, and PlantDoc datasets, respectively, surpassing previous studies [33].

3. Methodology

This section delineates the rigorous approach employed to conduct an exhaustive review of machine vision systems and artificial intelligence (AI) algorithms for the identification and harvesting of agricultural products, with a particular focus on their transformative role in advancing sustainable Solanum lycopersicum (tomato) disease management for global food security.

3.1. Identification of Relevant Articles

The first step of our literature study entailed a comprehensive search of prominent scientific databases, including IEEE Xplore, ScienceDirect, Springer, and Google Scholar, to locate pertinent publications on machine vision systems and AI algorithms for the identification and harvesting of agricultural products. To refine our search, we utilized targeted keywords such as “tomato disease detection”, “Solanum lycopersicum disease classification”, “AI algorithm for tomato disease detection”, “tomato disease classification using deep learning” and “tomato disease classification using machine learning”. The search was restricted to papers published between 2014 and 2023.

3.2. Inclusion and Exclusion Criteria

The present review paper diligently adhered to stringent inclusion and exclusion criteria, thereby ensuring the utmost relevance and scientific rigor of the studies under consideration. In particular, the inclusion criteria necessitated that the papers encompass AI algorithms for the identification and classification of diseases in Solanum lycopersicum, be published in peer-reviewed scientific journals or conference proceedings and be written in English. Conversely, studies failing to meet the aforementioned criteria, i.e., those that did not pertain to AI algorithms for the detection and classification of tomato leaf diseases, were not published in peer-reviewed scientific journals or conference proceedings, or were not documented in the English language, were excluded from the analysis.

3.3. Screening and Data Extraction

To ensure the accuracy and completeness of our review, two independent reviewers evaluated the eligibility of the submitted articles in accordance with the inclusion and exclusion criteria. Any disagreements were resolved through extensive discussions between the reviewers. After screening, pertinent data were extracted from the selected articles using a predetermined data extraction form. The extracted data included the author’s name, publication year, study area, and the AI algorithms used for the detection and classification of Solanum lycopersicum (Tomato) diseases detection.

3.4. Data Analysis

The extracted data was subjected to extensive descriptive statistics to uncover patterns and trends in the use of AI algorithms for the detection and classification of Solanum lycopersicum (Tomato) diseases. A flowchart was created to summarize the many stages of the review process, including the number of articles identified, vetted, and included. We also employed various graphical tools, including bar charts to visually illustrate the distribution of various types of data gathering techniques employed for AI algorithms for the detection and classification of Solanum lycopersicum (Tomato) diseases.

3.5. Article Screening Summary

Our comprehensive search of scientific databases yielded a total of 1051 articles, out of which 137 were ultimately included in this review paper. Figure 1 depicts the flowchart of the review process, highlighting the various stages involved. Figure 2 presents a bar chart showing the number of articles referred for review and published in each year from 2014 to 2023. The chart reveals interesting trends, with the highest number of articles being published and referred for review in 2022 (41), while certain years, such as 2014, 2015, 2016, 2017, and 2018, had relatively low numbers of articles referred for review. Overall, the chart visually highlights the relationship between the number of articles published and referred for review and can help identify emerging trends over time.
Through our rigorous methodology, this review paper presents a comprehensive analysis of the use of AI algorithms for the detection and classification of Solanum lycopersicum (Tomato) diseases and management for global food security.

4. Details about Tomato Diseases

Tomatoes, an essential crop that feeds millions worldwide, are nevertheless prone to a variety of diseases that can significantly reduce yields and quality. In this section, we will discuss nine common tomato diseases, along with their causal agents and symptoms. Among these diseases, early blight is a fungal disease caused by Alternaria solani that attacks tomato plants, resulting in significant damage to the plant’s foliage, stems, and fruit [34]. Initially, early blight symptoms appear on the plant’s lower leaves as small, circular, dark brown or black spots with a target-like pattern in the center. However, as the disease progresses, the spots enlarge, and the affected leaves may turn yellow and eventually die. Additionally, stem cankers may form, and fruit may become infected and rot, causing even more harm [35].
Another fungal disease that can cause significant damage to the foliage, stems, and fruit of tomato plants is late blight. Phytophthora infestans is the causal agent of this disease, which affects not only tomatoes but also potatoes [36]. Water-soaked lesions on leaves, stems, and fruit are the first signs of late blight, which quickly grow, turn brown, and become covered in a white mold. In severe cases, the entire plant can become infected and wilt. Late blight thrives in cool, moist environments with temperatures ranging from 50 to 70 °F and high humidity, and disease development is aided by prolonged leaf wetness and heavy dew [35].
Tomato septoria leaf spot is another fungal disease that can cause significant foliage damage to tomato plants. Septoria lycopersici is the causal agent of this disease, which initially appears as small, circular, brown or black spots on the plant’s lower leaves. As the disease progresses, the spots enlarge, and the center may fall out, giving the leaf a shot-hole appearance. In severe cases, the entire plant can become infected, with yellowing and dropping leaves [37]. Septoria leaf spot is most common in areas with frequent rainfall or overhead irrigation where cool and moist conditions favor disease development. High humidity also plays a contributing role [35].
Bacterial spot, caused by Xanthomonas campestris pv. vesicatoria, is a bacterial disease that can affect the foliage, stems, and fruit of tomato plants [38]. The first symptoms of bacterial spot appear as small, water-soaked, dark brown to black spots on the leaves, stems, and fruit of the plant. The spots may become surrounded by a yellow halo, and in many cases, the whole plant can become infected and defoliate. This disease thrives in warm, moist conditions, and is most common in areas with frequent rainfall or overhead irrigation. High humidity also contributes to disease development [35].
Tomato mosaic virus (ToMV) is a viral disease that causes yellow mottling, stunting, and malformation of tomato leaves, stems, and fruit. The severity of ToMV symptoms may vary depending on the virus strain and the stage of the plant when infected [39]. Not only does ToMV result in common signs such as yellow mottling, stunting, and deformation of leaves, stems, and fruit, but it can also cause curled, crinkled, and deformed leaves. The spread of ToMV is facilitated by a variety of insect vectors, including thrips, aphids, and whiteflies. Large numbers of these insects can lead to greater viral transmission [35].
Tomato yellow leaf curl virus (TYLCV) is another viral disease that can have detrimental effects on tomato plants. TYLCV causes yellowing, curling, and stunting of tomato leaves, as well as reduced yields and fruit quality. In fact, the symptoms of TYLCV include not only yellowing, curling, and stunting of leaves but also the thickening and leathery texture of leaves. The sweet potato whitefly (Bemisia tabaci) spreads TYLCV, and large numbers of these insects can contribute to its greater viral transmission [40].
Tomato leaf mold, on the other hand, is a fungal disease that can lead to yellowing, withering, and defoliation of tomato leaves, resulting in lower yields and fruit quality. In addition, a grayish-white mold may also develop on the leaves. High humidity and temperatures ranging from 60 to 80 °F favor the growth of leaf mold, which can become more severe in greenhouse or high-tunnel production methods [41].
The two-spotted spider mite (Tetranychus urticae) is a common tomato plant pest that can cause leaf yellowing, stippling, and mortality, as well as lower yields and fruit quality. Two-spotted spider mite damage can also result in the formation of webs on the plant’s leaves and stems [42]. Hot and dry weather conditions are ideal for two-spotted spider mites, which can grow faster in greenhouse or high-tunnel production systems with highly regulated environments [35].
Finally, tomato target spot is a fungal disease that causes circular, necrotic patches on the stems, leaves, and fruits of tomato plants. This disease can induce defoliation and yield reductions and may become more severe in locations with significant rains, particularly in warm and damp circumstances [35]. The symptoms of tomato target spot include the growth of circular, necrotic patches on the fruit, stems, and leaves, which may be encircled by a yellow halo [43]. Table 2 summarizes the common tomato diseases.

5. Datasets for Tomato Leaf Diseases

Numerous datasets have been developed to aid in the classification of tomato leaf diseases through the use of machine learning (ML) and deep learning (DL) algorithms. One notable dataset used for tomato leaf disease classification is the PlantVillage dataset, which boasts over 50,000 images of healthy and damaged plant leaves, including tomato leaves. This dataset was made available to the public by Hughes and Salathe (2015) and has been widely utilized in the field [44].
In addition, another dataset that has been developed for visually detecting plant diseases is PlantDoc. Singh et al. (2019) created this dataset by annotating images scraped from the internet, investing roughly 300 human hours to annotate 2598 data points spanning 13 plant species and 17 disease classifications. These datasets serve as excellent resources for researchers and professionals seeking to develop and test ML and DL models for accurate and efficient tomato leaf disease classification [45].
Table 3 provides a comprehensive list of such datasets that are available for tomato leaf disease classification, further emphasizing the wealth of resources available to researchers in this field.

6. Algorithms Used for Classification

In agriculture, the classification of tomato leaf diseases is not just important, but rather a crucial task since early detection and management can significantly enhance crop production and quality. To achieve this, both Machine Learning (ML) and Deep Learning (DL) methods have been extensively used, and this section covers some of the current research and their performance.
One study by Mohanty et al. (2016) trained a deep Convolutional Neural Network (CNN) to detect 14 crop species and 26 diseases [7]. Two architectures, namely LeNet-5 and AlexNet, were used, both of which adhere to the same design principles [6,128]. While LeNet-5 typically consists of one or more fully linked layers, stacked convolution layers, and one or more further layers, AlexNet has five convolutional layers followed by three fully connected layers and a SoftMax layer. In contrast, GoogLeNet has 22 layers and nine inception modules, which makes it deeper and more complex than the other two. The researchers used parallel processing to collect a variety of features simultaneously, and the outputs of all these parallel layers were concatenated using a filter concatenation layer [5,6].
Fuentes et al. (2017) demonstrated a DL-based technique for identifying diseases and pests in tomato plants using the Faster R-CNN model, which has two stages [8]. In the first stage, a feature extractor processes the image and feeds it into a Region Proposal Network (RPN), which scores each object proposition based on intermediate-level characteristics [129]. In the second stage, the system crops features from the feature map and feeds them to successive layers of the feature extractor in order to determine the class probability and bounding box for each proposal. Moreover, R-FCN and SSD meta-architectures were also used to address translation invariance and object recognition problems [10,130].
To tackle false positives and class imbalance, Fuentes et al. (2018) developed, “Refinement Filter Bank architecture for Tomato Plant Diseases and Pests Recognition” [18]. This architecture consists of a primary diagnostic unit, a secondary diagnostic unit, and an integration unit [71]. The primary unit generates a collection of bounding boxes, while the secondary unit filters each box’s confidence by training CNN classifiers, and the integration unit combines the outcomes from the main and secondary units [131].
Furthermore, in a separate study conducted by Tian et al. (2019), an enhanced K-means algorithm utilizing adaptive clustering numbers was proposed for the purpose of segmenting tomato leaf images. The researchers conducted a thorough investigation of various clustering numbers in order to identify the optimal segmentation, employing the Davies-Bouldin (DB) validity index as a metric for evaluation. Moreover, the excess green feature was employed as a discriminative factor to automatically distinguish the leaf from the background in the image [125,132].
Liu et al. (2020) proposed an enhanced YOLOv3 algorithm for detecting tomato diseases and pests. To improve the YOLOv3 network (Figure 3a), they incorporated object bounding box dimension clustering, multi-scale training, and multi-scale feature identification based on picture pyramids [133]. With a detection accuracy of 92.39% and a detection time of less than 20.39 milliseconds, the enhanced YOLOv3 algorithm exhibits exemplary performance. Therefore, it can be considered an optimal solution for prompt and efficient detection of tomato diseases and pests [126].
Upon conducting extensive review studies, it was found that ResNet50 (Figure 3b), VGG16 (Figure 4a), and AlexNet (Figure 4b) were the most commonly utilized algorithms for tomato leaf disease classification [6,16,29]. Table 4 provides a summary of the algorithms with the highest accuracies for the classification of tomato leaf diseases.

7. Features Used by Algorithms

ML and DL algorithms can employ a variety of features to classify tomato diseases. This section discusses several features described in the papers evaluated.

7.1. Texture Features

In order to effectively classify images using texture analysis, Basavaiah et al. (2020) employed machine learning techniques to train the feature vector. Texture is a crucial aspect of human visual perception, and statistical texture techniques are widely employed for image analysis and comprehension. These techniques locate each pixel’s local features and calculate a set of parameters from their distributions. Numerous applications use this method, which considers the spatial distribution of grey values [27]. The method involves two steps, namely computing the Gray-Level Co-occurrence Matrix (GLCM) and calculating the texture characteristics based on the GLCM.
In addition, Basavaiah and Anthony (2020) classified images using the commonly used Local Binary Patterns (LBP) and Histogram of Oriented Gradients (HOG) visual descriptors in computer vision. It has been demonstrated that combining LBP and HOG descriptors substantially improves detection efficacy [23].

7.2. Color Features

Basavaiah et al. (2020) employed ML techniques to classify images based on color analysis. The distribution of colors in an image can be represented through a color histogram, which is a fundamental aspect of image processing. To create a color histogram, every color space can be utilized and divided into a suitable number of ranges, each with identical color values [134]. In addition, the color histogram can be presented as a smooth function that estimates the number of pixels in a given color space [23].

7.3. Shape Features

Wu et al. (2023) utilize ML algorithms and shape analysis for image classification. Shape features are calculated using Hu-moments, which are obtained from the weighted average of pixel intensities in an image, where the weighting is based on intensity rather than location. Image moments, which quantify the quantity of white regions or pixels in a binary image, are used to calculate the shape feature [135]. However, it is important to note that moment-based similarity alone may not be sufficient, as two images with the same moment values may still appear different. Therefore, finding stable moments is crucial in shape analysis. Central moments are utilized for form matching since they are independent by scale, translation, and rotation. Central moments determine moments, a set of seven numbers that are image-invariant. Scaling, translation, reflection, and rotation do not alter the first six moments, but picture reflection can change the seventh moment’s sign [23].

7.4. Deep Learning Features

Neural networks are capable of instantly extracting deep learning features, making them a common practice in image analysis. The stack ensemble DL feature extractor method, developed by Kaur et al. (2022), is based on the pre-trained model EfficientNetB7, and is employed in conjunction with a bottom-up Feature Pyramid Network (FPN) for efficient feature extraction [136]. The FPN allows for accurate representation of objects at various scales, providing users with both low-level and high-level capabilities at every level of the network [71]. A comprehensive list of features and their corresponding techniques from all the reviewed papers is provided in Table 5, showcasing the wealth of information obtained through this approach.

8. Discussion on Challenges and Trends

Tomato leaf disease classification is not only important but also a challenging task in plant pathology due to several factors such as the similarity of symptoms amongst diseases, changes in plant growth phases, and the presence of external noise. However, researchers have made significant progress in this field in recent years, and various trends in tomato leaf disease classification have emerged. These trends include the use of advanced machine learning techniques, such as convolutional neural networks (CNNs), which have shown remarkable performance in disease detection and classification [101]. Additionally, the development of new and more comprehensive datasets, combined with the use of transfer learning approaches, has led to more accurate and reliable classification models [90]. Furthermore, the integration of different modalities such as spectral and thermal imaging has provided a more holistic approach to disease diagnosis and classification, resulting in higher accuracy and robustness [137].

8.1. Challenges

8.1.1. Similarity of Symptoms

Classifying tomato leaf diseases accurately is a challenging task due to the similarity in symptomatology amongst diseases. This can lead to inaccurate diagnosis and management, which in turn may result in economic losses and yield reductions. For example, differentiating between early blight and bacterial spot, two common tomato diseases, can be difficult as they share symptoms such as dark spots on leaves. To overcome this issue, researchers have explored the use of machine learning (ML) approaches to increase classification accuracy [56]. Decision trees and Support Vector Machines (SVMs) are two widely used ML techniques to classify tomato leaf diseases. By analyzing various parameters such as leaf shape, color, texture, and size, these algorithms can identify the presence of a specific disease [71]. The ML models can learn to differentiate between diseases that display similar symptoms by training on a dataset of tomato leaves with known diseases [83].

8.1.2. Differences in Plant Growth Stages

Furthermore, external factors such as weather conditions and the presence of other plant diseases can add to the complexity of tomato disease classification. Hence, accurate diagnosis is crucial for successful disease control and management. With the advancement of deep learning techniques, several models have been developed to improve the classification accuracy of tomato diseases, including transfer learning and ensemble learning methods. Transfer learning allows for the transfer of knowledge from pre-trained models to new datasets, while ensemble learning combines multiple models to achieve higher accuracy. These approaches have shown significant improvements in classifying tomato diseases, leading to better disease management and higher crop yields.

8.1.3. Environmental Noise

In addition to the challenges posed by the similarity of symptoms and changes in plant growth phases, another significant difficulty in classifying tomato leaf diseases is the presence of ambient noise from various sources such as sunshine, shadows, and background clutter. To mitigate the impact of noise and enhance classification accuracy, researchers have explored different techniques, including image segmentation and background subtraction. Some researchers have even developed customized algorithms tailored specifically to address noise in image datasets, such as the novel combination of the Asymptotic Non-Local Means algorithm (ANLM) and the Multi-channel Automatic Orientation Recurrent Attention Network (MAORANet) [109]. These innovative approaches have shown promising results in reducing noise interference and improving the accuracy of tomato leaf disease classification.

8.1.4. Limited Dataset

Labelled data is a crucial component for effectively classifying tomato diseases; however, the lack of it is a significant challenge. Obtaining large datasets for machine learning is time-consuming, expensive, and labor-intensive, which may negatively impact the accuracy of classification models [18]. Therefore, field surveys and visual inspections are typically conducted to collect labelled data, requiring experts to carefully inspect leaves for disease symptoms before photographing, labelling, and saving them. Nevertheless, creating deep learning systems for classifying tomato diseases is extremely difficult due to the absence of labelled data, which is necessary for accurately learning and recognizing complex patterns and features [109]. Without sufficient data, the model may become overfitted or underfitted, leading to inaccurate disease classification. To tackle this issue, researchers have investigated various methods such as data augmentation, transfer learning, and active learning to improve the availability and quality of labelled data [118].

8.1.5. Imbalanced Data

Data imbalance is indeed a crucial issue in tomato leaf disease classification, as it can lead to biased classification models. Consequently, the misclassification of the minority class can result in disease spread and financial loss. Therefore, researchers have proposed various methods to address this issue. One such method is under sampling, which involves removing some of the majority class samples to balance the dataset. Another approach is oversampling, which duplicates the minority class samples to increase their representation in the dataset. Moreover, class weighting assigns different weights to each class in the dataset, with the minority class being assigned larger weights and the majority class assigned lower weights. These methods have been developed to tackle data imbalance and enhance the accuracy of classification models [18].

8.2. Trends

8.2.1. Deep Learning

The classification of tomato diseases has made significant strides in recent years, thanks to the advent of deep learning (DL) algorithms, with convolutional neural networks (CNNs) playing a pivotal role in this advancement. CNNs have garnered widespread recognition as a powerful method for disease classification, owing to their ability to automatically discern complex patterns in images and extract high-level information [51]. In various image recognition applications, including tomato leaf disease classification, CNNs have demonstrated distinct advantages over traditional machine learning techniques. Numerous studies have utilized CNNs to accurately classify tomato leaf diseases, outperforming conventional methods and achieving remarkable levels of accuracy [69].
The impact of CNNs’ ability to accurately classify diseases in tomato plants will be profound in agriculture, as it will facilitate rapid identification and treatment of diseased plants by farmers, thereby minimizing crop losses and maximizing yields. The utilization of CNNs in the classification of tomato diseases represents a significant development in the field, with substantial potential to enhance crop management practices and ensure food security [119]. In summary, the use of CNNs in the classification of tomato diseases holds immense promise and is poised to revolutionize crop management techniques in the agriculture industry.

8.2.2. Transfer Learning

Transfer learning, a widely used approach for tomato leaf disease classification, has demonstrated its effectiveness in recent years. By utilizing pre-trained deep learning models on large datasets and adapting them to specific objectives such as identifying tomato leaf diseases, this technique has been shown to significantly enhance classification accuracy [46,60]. The advantage of transfer learning lies in the fact that pre-trained models have already mastered the skill of identifying and extracting useful features from massive amounts of data. By fine-tuning the pre-trained algorithms on a smaller dataset of tomato leaf images, they can quickly learn to recognize specific patterns and features associated with different types of tomato leaf diseases [63,64]. Furthermore, transfer learning requires less training data, which is one of its key benefits. This is due to the ability of the pre-trained model to transfer its knowledge for recognizing a wide range of features and patterns from large datasets to the task of classifying tomato leaf diseases. Additionally, transfer learning can reduce the time and resources needed for model training, making it a more efficient strategy for disease classification [76,112].

8.2.3. Ensemble Learning

Ensemble learning has emerged as a key technique in the classification of tomato diseases, integrating many classification models to increase overall performance and avoid overfitting. By combining the results from various models and using their combined predictions, ensemble learning has been shown to enhance the accuracy and robustness of disease classification models. Furthermore, ensemble learning can address the issue of overfitting, which is a prevalent problem in machine learning. Bagging, boosting, and stacking are three ways to apply ensemble learning, each having specific advantages and disadvantages [31].
Using ensemble learning techniques can boost the precision and dependability of models used to classify tomato diseases, ultimately resulting in better crop management techniques and higher yields. Researchers and professionals can benefit from the power of ensemble learning to improve the detection and classification system’s accuracy and robustness, integrating various classification models such as decision trees, SVM, and neural networks.
Despite the challenges in tomato disease classification using ML and DL, including symptom similarities between diseases, changes in plant growth phases, and ambient noise, the field has made significant advancements in recent years [76]. These advancements include the development of novel deep learning architectures and the integration of different data modalities such as hyperspectral imaging and environmental data. Furthermore, the use of explainable AI methods to increase model interpretability and facilitate their adoption by end-users such as farmers and crop consultants is gaining traction [104].
In summary, tomato leaf disease classification using ML and DL holds great promise for improving crop management practices and reducing the impact of plant diseases on food production. These developments point to the possibility of building accurate and efficient systems for tomato leaf disease classification.

9. Conclusions

This review article sheds light on the current state of research on tomato disease classification, which is a critical aspect of ensuring sustainable food security. The availability of datasets, utilization of algorithms, and identification of the highest accuracy algorithms are all key takeaways from this review. Among the available datasets, the publicly accessible PlantVillage dataset stands out as the most popular for tomato leaf disease classification. ResNet50, VGG16, and AlexNet emerge as the most frequently utilized algorithms in this field, exhibiting remarkable accuracy rates. Despite the existing challenges, such as the need for high-quality datasets and improved methods for model interpretability, the application of machine learning (ML) and deep learning (DL) in disease classification holds immense promise for further advancements. This review paper provides a comprehensive overview of the diverse ML and DL models applied, dataset acquisition approaches, feature extraction techniques, and performance evaluation criteria, providing a holistic picture of the current status of the field. As a valuable resource, it serves as a guide for professionals and researchers interested in developing more accurate and efficient methods for classifying tomato leaf diseases, thereby contributing to sustainable food security (SDG-3).

Author Contributions

Conceptualization, B.S., S.J. and N.K.; methodology, B.S., S.J. and N.K.; software, B.S. and S.J.; formal analysis, B.S. and S.J.; data curation, N.K.; writing—original draft preparation, B.S. and S.J.; writing—review and editing, N.K.; supervision, N.K.; project administration, N.K. All authors have read and agreed to the published version of the manuscript.

Funding

This research received no external funding and The APC was funded by Dr. Narendra Khatri.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

All data generated or analyzed during the review are included in this article.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Li, L.; Zhang, S.; Wang, B. Plant Disease Detection and Classification by Deep Learning—A Review. IEEE Access 2021, 9, 56683–56698. [Google Scholar] [CrossRef]
  2. Hariharan, J.; Ampatzidis, Y.; Abdulridha, J.; Batuman, O. An AI-based spectral data analysis process for recognizing unique plant biomarkers and disease features. Comput. Electron. Agric. 2023, 204, 107574. [Google Scholar] [CrossRef]
  3. Sahu, S.K.; Pandey, M. An optimal hybrid multiclass SVM for plant leaf disease detection using spatial Fuzzy C-Means model. Expert Syst. Appl. 2023, 214, 118989. [Google Scholar] [CrossRef]
  4. Ahmed, I.; Yadav, P.K. A systematic analysis of machine learning and deep learning based approaches for identifying and diagnosing plant diseases. Sustain. Oper. Comput. 2023, 4, 96–104. [Google Scholar] [CrossRef]
  5. Szegedy, C.; Liu, W.; Jia, Y.; Sermanet, P.; Reed, S.; Anguelov, D.; Erhan, D.; Vanhoucke, V.; Rabinovich, A.; Liu, W.; et al. Going deeper with convolutions. In Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 7–12 June 2015; pp. 1–9. [Google Scholar] [CrossRef] [Green Version]
  6. Krizhevsky, A.; Sutskever, I.; Hinton, G.E. Imagenet classification with deep convolutional neural networks. Commun. ACM 2017, 60, 84–90. [Google Scholar] [CrossRef] [Green Version]
  7. Mohanty, S.P.; Hughes, D.P.; Salathé, M. Using Deep Learning for Image-Based Plant Disease Detection. Front. Plant Sci. 2016, 7, 1419. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  8. Fuentes, A.; Yoon, S.; Kim, C.S.; Park, S.D. A Robust Deep-Learning-Based Detector for Real-Time Tomato Plant Diseases and Pests Recognition. Precis. Agric. 2017, 17, 2022. [Google Scholar] [CrossRef] [Green Version]
  9. Ren, S.; He, K.; Girshick, R.; Sun, J. Faster R-CNN: Towards real-time object detection with region proposal networks. IEEE Trans. Pattern Anal. Mach. Intell. 2017, 39, 1137–1149. [Google Scholar] [CrossRef] [Green Version]
  10. Dai, J.; Li, Y.; He, K.; Sun, J. R-FCN: Object Detection via Region-based Fully Convolutional Networks. In Advances in Neural Information Processing Systems 29 (NIPS 2016); Curran Associates Inc.: Red Hook, NY, USA, 2016; pp. 379–387. [Google Scholar]
  11. Li, M.; Zhou, G.; Chen, A.; Li, L.; Hu, Y. Identification of tomato leaf diseases based on LMBRNet. Eng. Appl. Artif. Intell. 2023, 123, 106195. [Google Scholar] [CrossRef]
  12. Sunil, C.K.; Jaidhar, C.D.; Patil, N. Tomato plant disease classification using Multilevel Feature Fusion with adaptive channel spatial and pixel attention mechanism. Expert Syst. Appl. 2023, 228, 120381. [Google Scholar] [CrossRef]
  13. Sanida, T.; Sideris, A.; Sanida, M.V.; Dasygenis, M. Tomato leaf disease identification via two–stage transfer learning approach. Smart Agric. Technol. 2023, 5, 100275. [Google Scholar] [CrossRef]
  14. Huang, J.; Rathod, V.; Sun, C.; Zhu, M.; Korattikara, A.; Fathi, A.; Fischer, I.; Wojna, Z.; Song, Y.; Guadarrama, S.; et al. Speed/accuracy trade-offs for modern convolutional object detectors. In Proceedings of the 30th IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 July 2016; pp. 3296–3305. [Google Scholar]
  15. D’amore, G.; Di Vaio, A.; Balsalobre-Lorente, D.; Boccia, F. Artificial Intelligence in the Water–Energy–Food Model: A Holistic Approach towards Sustainable Development Goals. Sustainability 2022, 14, 867. [Google Scholar] [CrossRef]
  16. Simonyan, K.; Zisserman, A. Very Deep Convolutional Networks for Large-Scale Image Recognition. arXiv 2014, arXiv:1409.1556. [Google Scholar]
  17. Ferentinos, K.P. Deep learning models for plant disease detection and diagnosis. Comput. Electron. Agric. 2018, 145, 311–318. [Google Scholar] [CrossRef]
  18. Fuentes, A.F.; Yoon, S.; Lee, J.; Park, D.S. High-Performance Deep Neural Network-Based Tomato Plant Diseases and Pests Diagnosis System with Refinement Filter Bank. Front. Plant Sci. 2018, 9, 1162. [Google Scholar] [CrossRef] [Green Version]
  19. Kumar, S.K.P.; Sumithra, M.G.; Saranya, N. Artificial bee colony-based fuzzy c means (ABC-FCM) segmentation algorithm and dimensionality reduction for leaf disease detection in bioinformatics. J. Supercomput. 2019, 75, 8293–8311. [Google Scholar] [CrossRef]
  20. Howard, A.G.; Zhu, M.; Chen, B.; Kalenichenko, D.; Wang, W.; Weyand, T.; Andreetto, M.; Adam, H. MobileNets: Efficient Convolutional Neural Networks for Mobile Vision Applications. arXiv 2017, arXiv:1704.04861. [Google Scholar]
  21. Kc, K.; Yin, Z.; Wu, M.; Wu, Z. Depthwise separable convolution architectures for plant disease classification. Comput. Electron. Agric. 2019, 165, 104948. [Google Scholar] [CrossRef]
  22. Geetharamani, G.; Arun Pandian, J. Identification of plant leaf diseases using a nine-layer deep convolutional neural network. Comput. Electr. Eng. 2019, 76, 323–338. [Google Scholar] [CrossRef]
  23. Basavaiah, J.; Anthony, A.A. Tomato Leaf Disease Classification using Multiple Feature Extraction Techniques. Wirel. Pers. Commun. 2020, 115, 633–651. [Google Scholar] [CrossRef]
  24. Chen, X.; Zhou, G.; Chen, A.; Yi, J.; Zhang, W.; Hu, Y. Identification of tomato leaf diseases based on combination of ABCK-BWTR and B-ARNet. Comput. Electron. Agric. 2020, 178, 105730. [Google Scholar] [CrossRef]
  25. Karthik, R.; Hariharan, M.; Anand, S.; Mathikshara, P.; Johnson, A.; Menaka, R. Attention embedded residual CNN for disease detection in tomato leaves. Appl. Soft Comput. 2019, 86, 105933. [Google Scholar] [CrossRef]
  26. Li, Z.; Yang, Y.; Li, Y.; Guo, R.; Yang, J.; Yue, J. A solanaceae disease recognition model based on SE-Inception. Comput. Electron. Agric. 2020, 178, 105792. [Google Scholar] [CrossRef]
  27. Tan, L.; Lu, J.; Jiang, H. Tomato Leaf Diseases Classification Based on Leaf Images: A Comparison between Classical Machine Learning and Deep Learning Methods. Agriengineering 2021, 3, 542–558. [Google Scholar] [CrossRef]
  28. Tan, M.; Le, Q.V. EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks. arXiv 2019, arXiv:1905.11946. [Google Scholar]
  29. He, K.; Zhang, X.; Ren, S.; Sun, J. Deep residual learning for image recognition. In Proceedings of the IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA, 27–30 June 2016; pp. 770–778. [Google Scholar] [CrossRef] [Green Version]
  30. Sandler, M.; Howard, A.; Zhu, M.; Zhmoginov, A.; Chen, L. MobileNetV2: Inverted Residuals and Linear Bottlenecks. In Proceedings of the 2018 IEEE Conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA, 18–22 June 2018; pp. 4510–4520. [Google Scholar] [CrossRef] [Green Version]
  31. Peker, M. Multi-channel capsule network ensemble for plant disease detection. SN Appl. Sci. 2021, 3, 707. [Google Scholar] [CrossRef]
  32. Wang, D.; Wang, J.; Ren, Z.; Li, W. DHBP: A dual-stream hierarchical bilinear pooling model for plant disease multi-task classification. Comput. Electron. Agric. 2022, 195, 106788. [Google Scholar] [CrossRef]
  33. Yu, S.; Xie, L.; Huang, Q. Inception convolutional vision transformers for plant disease identification. Internet Things 2023, 21, 100650. [Google Scholar] [CrossRef]
  34. Roy, C.K.; Akter, N.; Sarkar, M.K.; Pk, M.U.; Begum, N.; Zenat, E.A.; Jahan, M.A. Control of Early Blight of Tomato Caused by Alternaria solani and Screening of Tomato Varieties against the Pathogen. Open Microbiol. J. 2019, 13, 41–50. [Google Scholar] [CrossRef]
  35. Jones, J.B.; Zitter, T.A.; Momol, T.M.; Miller, S.A. PART I: Infectious Diseases. In Compendium of Tomato Diseases and Pests, 2nd ed.; APS Publications: St. Paul, MN, USA, 2016. [Google Scholar] [CrossRef]
  36. Mulugeta, T.; Abreha, K.; Tekie, H.; Mulatu, B.; Yesuf, M.; Andreasson, E.; Liljeroth, E.; Alexandersson, E. Phosphite protects against potato and tomato late blight in tropical climates and has varying toxicity depending on the Phytophthora infestans isolate. Crop. Prot. 2019, 121, 139–146. [Google Scholar] [CrossRef]
  37. Summers, C.F.; Park, S.; Dunn, A.R.; Rong, X.; Everts, K.L.; Kleinhenz, M.D.; Gardener, B.M.; Smart, C.D. Fungal and oomycete pathogen detection in the rhizosphere of organic tomatoes grown in cover crop-treated soils. Appl. Soil Ecol. 2014, 80, 44–50. [Google Scholar] [CrossRef]
  38. Príncipe, A.; Fernandez, M.; Torasso, M.; Godino, A.; Fischer, S. Effectiveness of tailocins produced by Pseudomonas fluorescens SF4c in controlling the bacterial-spot disease in tomatoes caused by Xanthomonas vesicatoria. Microbiol. Res. 2018, 212–213, 94–102. [Google Scholar] [CrossRef] [PubMed]
  39. Pozharskiy, A.; Kostyukova, V.; Taskuzhina, A.; Nizamdinova, G.; Kisselyova, N.; Kalendar, R.; Karimov, N.; Gritsenko, D. Screening a collection of local and foreign varieties of Solanum lycopersicum L. in Kazakhstan for genetic markers of resistance against three tomato viruses. Heliyon 2022, 8, e10095. [Google Scholar] [CrossRef] [PubMed]
  40. Sade, D.; Sade, N.; Brotman, Y.; Czosnek, H. Tomato yellow leaf curl virus (TYLCV)-resistant tomatoes share molecular mechanisms sustaining resistance with their wild progenitor Solanum habrochaites but not with TYLCV-susceptible tomatoes. Plant Sci. 2020, 295, 110439. [Google Scholar] [CrossRef]
  41. Zhao, T.; Pei, T.; Jiang, J.; Yang, H.; Zhang, H.; Li, J.; Xu, X. Understanding the mechanisms of resistance to tomato leaf mold: A review. Hortic. Plant J. 2022, 8, 667–675. [Google Scholar] [CrossRef]
  42. Osman, M.A.; Al Dhafar, Z.M.; Alqahtani, A.M. Biological responses of the two-spotted spider mite, Tetranychus urticae to different host plant. Arch. Phytopathol. Plant Prot. 2019, 52, 1229–1238. [Google Scholar] [CrossRef]
  43. Sahu, P.K.; Singh, S.; Gupta, A.; Singh, U.B.; Brahmaprakash, G.; Saxena, A.K. Antagonistic potential of bacterial endophytes and induction of systemic resistance against collar rot pathogen Sclerotium rolfsii in tomato. Biol. Control 2019, 137, 104014. [Google Scholar] [CrossRef]
  44. Hughes, D.P.; Salathe, M. An open access repository of images on plant health to enable the development of mobile disease diagnostics. arXiv 2015, arXiv:1511.08060. [Google Scholar]
  45. Singh, D.; Jain, N.; Jain, P.; Kayal, P.; Kumawat, S.; Batra, N. PlantDoc: A Dataset for Visual Plant Disease Detection. arXiv 2019, arXiv:1911.10317. [Google Scholar] [CrossRef] [Green Version]
  46. Abbas, A.; Jain, S.; Gour, M.; Vankudothu, S. Tomato plant disease detection using transfer learning with C-GAN synthetic images. Comput. Electron. Agric. 2021, 187, 106279. [Google Scholar] [CrossRef]
  47. Albahli, S.; Nawaz, M. DCNet: DenseNet-77-based CornerNet model for the tomato plant leaf disease detection and classification. Front. Plant Sci. 2022, 13, 957961. [Google Scholar] [CrossRef]
  48. Vadivel, T.; Suguna, R. Automatic recognition of tomato leaf disease using fast enhanced learning with image processing. Acta Agric. Scand. Sect. B Soil Plant Sci. 2021, 72, 312–324. [Google Scholar] [CrossRef]
  49. Zhao, S.; Peng, Y.; Liu, J.; Wu, S. Tomato Leaf Disease Diagnosis Based on Improved Convolution Neural Network by Attention Module. Agriculture 2021, 11, 651. [Google Scholar] [CrossRef]
  50. Zhao, Y.; Sun, C.; Xu, X.; Chen, J. RIC-Net: A plant disease classification model based on the fusion of Inception and residual structure and embedded attention mechanism. Comput. Electron. Agric. 2022, 193, 106644. [Google Scholar] [CrossRef]
  51. Nandhini, S.; Ashokkumar, K. Improved crossover based monarch butterfly optimization for tomato leaf disease classification using convolutional neural network. Multimed. Tools Appl. 2021, 80, 18583–18610. [Google Scholar] [CrossRef]
  52. Gadekallu, T.R.; Rajput, D.S.; Reddy, M.P.K.; Lakshmanna, K.; Bhattacharya, S.; Singh, S.; Jolfaei, A.; Alazab, M. A novel PCA–whale optimization-based deep neural network model for classification of tomato plant diseases using GPU. J. Real-Time Image Process. 2020, 18, 1383–1396. [Google Scholar] [CrossRef]
  53. Wspanialy, P.; Moussa, M. A detection and severity estimation system for generic diseases of tomato greenhouse plants. Comput. Electron. Agric. 2020, 178, 105701. [Google Scholar] [CrossRef]
  54. Hassan, S.M.; Maji, A.K. Deep feature-based plant disease identification using machine learning classifier. Innov. Syst. Softw. Eng. 2022, 1–11. [Google Scholar] [CrossRef]
  55. Mondal, D.; Roy, K.; Pal, D.; Kole, D.K. Deep Learning-Based Approach to Detect and Classify Signs of Crop Leaf Diseases and Pest Damage. SN Comput. Sci. 2022, 3, 433. [Google Scholar] [CrossRef]
  56. Lamba, S.; Saini, P.; Kaur, J.; Kukreja, V. Optimized classification model for plant diseases using generative adversarial networks. Innov. Syst. Softw. Eng. 2023, 19, 103–115. [Google Scholar] [CrossRef]
  57. Aishwarya, N.; Praveena, N.G.; Priyanka, S.; Pramod, J. Smart farming for detection and identification of tomato plant diseases using light weight deep neural network. Multimed. Tools Appl. 2022, 82, 18799–18810. [Google Scholar] [CrossRef]
  58. Russel, N.S.; Selvaraj, A. Leaf species and disease classification using multiscale parallel deep CNN architecture. Neural Comput. Appl. 2022, 34, 19217–19237. [Google Scholar] [CrossRef]
  59. Moussafir, M.; Chaibi, H.; Saadane, R.; Chehri, A.; El Rharras, A.; Jeon, G. Design of efficient techniques for tomato leaf disease detection using genetic algorithm-based and deep neural networks. Plant Soil 2022, 479, 251–266. [Google Scholar] [CrossRef]
  60. Liu, G.; Peng, J.; El-Latif, A.A.A. SK-MobileNet: A Lightweight Adaptive Network Based on Complex Deep Transfer Learning for Plant Disease Recognition. Arab. J. Sci. Eng. 2022, 48, 1661–1675. [Google Scholar] [CrossRef]
  61. Mahmoud, M.A.B.; Guo, P.; Wang, K. Pseudoinverse learning autoencoder with DCGAN for plant diseases classification. Multimed. Tools Appl. 2020, 79, 26245–26263. [Google Scholar] [CrossRef]
  62. Agarwal, M.; Gupta, S.K.; Biswas, K. Development of Efficient CNN model for Tomato crop disease identification. Sustain. Comput. Inform. Syst. 2020, 28, 100407. [Google Scholar] [CrossRef]
  63. Shang, F.; Muthanna, M.S.A.; Khayyat, M.; Al-Gaashani, M.S.A.M.; El-Latif, A.A.A. Tomato leaf disease classification by exploiting transfer learning and feature concatenation. IET Image Process. 2022, 16, 913–925. [Google Scholar] [CrossRef]
  64. Thangaraj, R.; Anandamurugan, S.; Kaliappan, V.K. Automated tomato leaf disease classification using transfer learning-based deep convolution neural network. J. Plant Dis. Prot. 2020, 128, 73–86. [Google Scholar] [CrossRef]
  65. Gonzalez-Huitron, V.; León-Borges, J.A.; Rodriguez-Mata, A.; Amabilis-Sosa, L.E.; Ramírez-Pereda, B.; Rodriguez, H. Disease detection in tomato leaves via CNN with lightweight architectures implemented in Raspberry Pi 4. Comput. Electron. Agric. 2021, 181, 105951. [Google Scholar] [CrossRef]
  66. Maeda-Gutiérrez, V.; Galván-Tejada, C.E.; Zanella-Calzada, L.A.; Celaya-Padilla, J.M.; Galván-Tejada, J.I.; Gamboa-Rosales, H.; Luna-García, H.; Magallanes-Quintanar, R.; Guerrero Méndez, C.A.; Olvera-Olvera, C.A. Comparison of Convolutional Neural Network Architectures for Classification of Tomato Plant Diseases. Appl. Sci. 2020, 10, 1245. [Google Scholar] [CrossRef] [Green Version]
  67. Brahimi, M.; Boukhalfa, K.; Moussaoui, A. Deep Learning for Tomato Diseases: Classification and Symptoms Visualization. Appl. Artif. Intell. 2017, 31, 299–315. [Google Scholar] [CrossRef]
  68. Li, M.; Zhou, G.; Chen, A.; Yi, J.; Lu, C.; He, M.; Hu, Y. FWDGAN-based data augmentation for tomato leaf disease identification. Comput. Electron. Agric. 2022, 194, 106779. [Google Scholar] [CrossRef]
  69. Anandhakrishnan, T.; Jaisakthi, S. Deep Convolutional Neural Networks for image based tomato leaf disease detection. Sustain. Chem. Pharm. 2022, 30, 100793. [Google Scholar] [CrossRef]
  70. Shoaib, M.; Hussain, T.; Shah, B.; Ullah, I.; Shah, S.M.; Ali, F.; Park, S.H. Deep learning-based segmentation and classification of leaf images for detection of tomato plant disease. Front. Plant Sci. 2022, 13, 1031748. [Google Scholar] [CrossRef]
  71. Kaur, P.; Harnal, S.; Gautam, V.; Singh, M.P.; Singh, S.P. An approach for characterization of infected area in tomato leaf disease based on deep learning and object detection technique. Eng. Appl. Artif. Intell. 2022, 115, 105210. [Google Scholar] [CrossRef]
  72. Hanh, B.T.; Van Manh, H.; Nguyen, N.-V. Enhancing the performance of transferred efficientnet models in leaf image-based plant disease classification. J. Plant Dis. Prot. 2022, 129, 623–634. [Google Scholar] [CrossRef]
  73. Bhagat, M.; Kumar, D. Performance evaluation of PCA based reduced features of leaf images extracted by DWT using random Forest and XGBoost classifier. Multimed. Tools Appl. 2023, 82, 26225–26254. [Google Scholar] [CrossRef]
  74. Ghofrani, A.; Toroghi, R.M. Knowledge distillation in plant disease recognition. Neural Comput. Appl. 2022, 34, 14287–14296. [Google Scholar] [CrossRef]
  75. Li, Y.; Chao, X. Semi-supervised few-shot learning approach for plant diseases recognition. Plant Methods 2021, 17, 68. [Google Scholar] [CrossRef]
  76. Vallabhajosyula, S.; Sistla, V.; Kolli, V.K.K. Transfer learning-based deep ensemble neural network for plant leaf disease detection. J. Plant Dis. Prot. 2021, 129, 545–558. [Google Scholar] [CrossRef]
  77. Albattah, W.; Nawaz, M.; Javed, A.; Masood, M.; Albahli, S. A novel deep learning method for detection and classification of plant diseases. Complex Intell. Syst. 2021, 8, 507–524. [Google Scholar] [CrossRef]
  78. Nawaz, M.; Nazir, T.; Javed, A.; Masood, M.; Rashid, J.; Kim, J.; Hussain, A. A robust deep learning approach for tomato plant leaf disease localization and classification. Sci. Rep. 2022, 12, 18568. [Google Scholar] [CrossRef] [PubMed]
  79. Sanida, T.; Tsiktsiris, D.; Sideris, A.; Dasygenis, M. A heterogeneous implementation for plant disease identification using deep learning. Multimed. Tools Appl. 2022, 81, 15041–15059. [Google Scholar] [CrossRef]
  80. Indu, V.T.; Priyadharsini, S.S. Crossover-based wind-driven optimized convolutional neural network model for tomato leaf disease classification. J. Plant Dis. Prot. 2021, 129, 559–578. [Google Scholar] [CrossRef]
  81. Tuncer, A. Cost-optimized hybrid convolutional neural networks for detection of plant leaf diseases. J. Ambient. Intell. Humaniz. Comput. 2021, 12, 8625–8636. [Google Scholar] [CrossRef]
  82. Pattanaik, P.A.; Khan, M.Z.; Patnaik, P.K. ILCAN: A New Vision Attention-Based Late Blight Disease Localization and Classification. Arab. J. Sci. Eng. 2021, 47, 2305–2314. [Google Scholar] [CrossRef]
  83. Prabu, M.; Chelliah, B.J. An intelligent approach using boosted support vector machine based arithmetic optimization algorithm for accurate detection of plant leaf disease. Pattern Anal. Appl. 2022, 26, 367–379. [Google Scholar] [CrossRef]
  84. Parasuraman, K.; Srinivasan, R.; Karunagaran, S. A novel technique of leaf disease classification with encryption techniques for secured cloud data service. J. Ambient. Intell. Humaniz. Comput. 2021, 1–17. [Google Scholar] [CrossRef]
  85. Salmi, A.; Benierbah, S.; Ghazi, M. Low complexity image enhancement GAN-based algorithm for improving low-resolution image crop disease recognition and diagnosis. Multimed. Tools Appl. 2022, 81, 8519–8538. [Google Scholar] [CrossRef]
  86. Nagi, R.; Tripathy, S.S. Plant disease identification using fuzzy feature extraction and PNN. Signal Image Video Process. 2023, 17, 2809–2815. [Google Scholar] [CrossRef]
  87. Kurmi, Y.; Saxena, P.; Kirar, B.S.; Gangwar, S.; Chaurasia, V.; Goel, A. Deep CNN model for crops’ diseases detection using leaf images. Multidimens. Syst. Signal Process. 2022, 33, 981–1000. [Google Scholar] [CrossRef]
  88. Kurmi, Y.; Gangwar, S.; Chaurasia, V.; Goel, A. Leaf images classification for the crops diseases detection. Multimed. Tools Appl. 2022, 81, 8155–8178. [Google Scholar] [CrossRef]
  89. Kurmi, Y.; Gangwar, S.; Agrawal, D.; Kumar, S.; Srivastava, H.S. Leaf image analysis-based crop diseases classification. Signal Image Video Process. 2020, 15, 589–597. [Google Scholar] [CrossRef]
  90. Mishra, M.; Choudhury, P.; Pati, B. Modified ride-NN optimizer for the IoT based plant disease detection. J. Ambient. Intell. Humaniz. Comput. 2020, 12, 691–703. [Google Scholar] [CrossRef]
  91. Alguliyev, R.; Imamverdiyev, Y.; Sukhostat, L.; Bayramov, R. Plant disease detection based on a deep model. Soft Comput. 2021, 25, 13229–13242. [Google Scholar] [CrossRef]
  92. Falaschetti, L.; Manoni, L.; Di Leo, D.; Pau, D.; Tomaselli, V.; Turchetti, C. A CNN-based image detector for plant leaf diseases classification. Hardwarex 2022, 12, e00363. [Google Scholar] [CrossRef]
  93. Kaya, Y.; Gürsoy, E. A novel multi-head CNN design to identify plant diseases using the fusion of RGB images. Ecol. Inform. 2023, 75, 101998. [Google Scholar] [CrossRef]
  94. Reddy, S.R.; Varma, G.S.; Davuluri, R.L. Resnet-based modified red deer optimization with DLCNN classifier for plant disease identification and classification. Comput. Electr. Eng. 2023, 105, 108492. [Google Scholar] [CrossRef]
  95. Atila, Ü.; Uçar, M.; Akyol, K.; Uçar, E. Plant leaf disease classification using EfficientNet deep learning model. Ecol. Inform. 2021, 61, 101182. [Google Scholar] [CrossRef]
  96. Pandey, A.; Jain, K. A robust deep attention dense convolutional neural network for plant leaf disease identification and classification from smart phone captured real world images. Ecol. Inform. 2022, 70, 101725. [Google Scholar] [CrossRef]
  97. Zhang, L.; Zhou, G.; Lu, C.; Chen, A.; Wang, Y.; Li, L.; Cai, W. MMDGAN: A fusion data augmentation method for tomato-leaf disease identification. Appl. Soft Comput. 2022, 123, 108969. [Google Scholar] [CrossRef]
  98. Wei, S.J.; Al Riza, D.F.; Nugroho, H. Comparative study on the performance of deep learning implementation in the edge computing: Case study on the plant leaf disease identification. J. Agric. Food Res. 2022, 10, 100389. [Google Scholar] [CrossRef]
  99. Argüeso, D.; Picon, A.; Irusta, U.; Medela, A.; San-Emeterio, M.G.; Bereciartua, A.; Alvarez-Gila, A. Few-Shot Learning approach for plant disease classification using images taken in the field. Comput. Electron. Agric. 2020, 175, 105542. [Google Scholar] [CrossRef]
  100. Gokulnath, B.V.; Usha Devi, G. Identifying and classifying plant disease using resilient LF-CNN. Ecol. Inform. 2021, 63, 101283. [Google Scholar] [CrossRef]
  101. Zhang, Y.; Song, C.; Zhang, D. Deep Learning-Based Object Detection Improvement for Tomato Disease. IEEE Access 2020, 8, 56607–56614. [Google Scholar] [CrossRef]
  102. Zhou, C.; Zhou, S.; Xing, J.; Song, J. Tomato Leaf Disease Identification by Restructured Deep Residual Dense Network. IEEE Access 2021, 9, 28822–28831. [Google Scholar] [CrossRef]
  103. Patil, M.A.; Manohar, M. Enhanced radial basis function neural network for tomato plant disease leaf image segmentation. Ecol. Inform. 2022, 70, 101752. [Google Scholar] [CrossRef]
  104. Astani, M.; Hasheminejad, M.; Vaghefi, M. A diverse ensemble classifier for tomato disease recognition. Comput. Electron. Agric. 2022, 198, 107054. [Google Scholar] [CrossRef]
  105. Xiang, S.; Liang, Q.; Sun, W.; Zhang, D.; Wang, Y. L-CSMS: Novel lightweight network for plant disease severity recognition. J. Plant Dis. Prot. 2021, 128, 557–569. [Google Scholar] [CrossRef]
  106. Cristin, R.; Kumar, B.S.; Priya, C.; Karthick, K. Deep neural network based Rider-Cuckoo Search Algorithm for plant disease detection. Artif. Intell. Rev. 2020, 53, 4993–5018. [Google Scholar] [CrossRef]
  107. Gao, R.; Wang, R.; Feng, L.; Li, Q.; Wu, H. Dual-branch, efficient, channel attention-based crop disease identification. Comput. Electron. Agric. 2021, 190, 106410. [Google Scholar] [CrossRef]
  108. Wang, D.; Wang, J.; Li, W.; Guan, P. T-CNN: Trilinear convolutional neural networks model for visual detection of plant diseases. Comput. Electron. Agric. 2021, 190, 106468. [Google Scholar] [CrossRef]
  109. Zhang, Y.; Huang, S.; Zhou, G.; Hu, Y.; Li, L. Identification of tomato leaf diseases based on multi-channel automatic orientation recurrent attention network. Comput. Electron. Agric. 2023, 205, 107605. [Google Scholar] [CrossRef]
  110. Chug, A.; Bhatia, A.; Singh, A.P.; Singh, D. A novel framework for image-based plant disease detection using hybrid deep learning approach. Soft Comput. 2022, 1–26. [Google Scholar] [CrossRef]
  111. Gehlot, M.; Gandhi, G.C. “EffiNet-TS”: A deep interpretable architecture using EfficientNet for plant disease detection and visualization. J. Plant Dis. Prot. 2023, 130, 413–430. [Google Scholar] [CrossRef]
  112. Kaur, P.; Harnal, S.; Gautam, V.; Singh, M.P.; Singh, S.P. A novel transfer deep learning method for detection and classification of plant leaf disease. J. Ambient. Intell. Humaniz. Comput. 2022, 1–18. [Google Scholar] [CrossRef]
  113. Ahmad, N.; Asif, H.M.S.; Saleem, G.; Younus, M.U.; Anwar, S.; Anjum, M.R. Leaf Image-Based Plant Disease Identification Using Color and Texture Features. Wirel. Pers. Commun. 2021, 121, 1139–1168. [Google Scholar] [CrossRef]
  114. Gajjar, R.; Gajjar, N.; Thakor, V.J.; Patel, N.P.; Ruparelia, S. Real-time detection and identification of plant leaf diseases using convolutional neural networks on an embedded platform. Vis. Comput. 2021, 38, 2923–2938. [Google Scholar] [CrossRef]
  115. Pardede, H.F.; Suryawati, E.; Zilvan, V.; Ramdan, A.; Kusumo, R.B.S.; Heryana, A.; Yuwana, R.S.; Krisnandi, D.; Subekti, A.; Fauziah, F.; et al. Plant diseases detection with low resolution data using nested skip connections. J. Big Data 2020, 7, 57. [Google Scholar] [CrossRef]
  116. Gui, P.; Dang, W.; Zhu, F.; Zhao, Q. Towards automatic field plant disease recognition. Comput. Electron. Agric. 2021, 191, 106523. [Google Scholar] [CrossRef]
  117. Chen, J.; Zhang, D.; Suzauddola; Zeb, A. Identifying crop diseases using attention embedded MobileNet-V2 model. Appl. Soft Comput. 2021, 113, 107901. [Google Scholar] [CrossRef]
  118. Huang, X.; Chen, A.; Zhou, G.; Zhang, X.; Wang, J.; Peng, N.; Yan, N.; Jiang, C. Tomato Leaf Disease Detection System Based on FC-SNDPN. Multimed. Tools Appl. 2022, 82, 2121–2144. [Google Scholar] [CrossRef]
  119. Cheng, H.-H.; Dai, Y.-L.; Lin, Y.; Hsu, H.-C.; Lin, C.-P.; Huang, J.-H.; Chen, S.-F.; Kuo, Y.-F. Identifying tomato leaf diseases under real field conditions using convolutional neural networks and a chatbot. Comput. Electron. Agric. 2022, 202, 107365. [Google Scholar] [CrossRef]
  120. Wang, C.; Zhou, J.; Zhang, Y.; Wu, H.; Zhao, C.; Teng, G.; Li, J. A Plant Disease Recognition Method Based on Fusion of Images and Graph Structure Text. Front. Plant Sci. 2022, 12, 731688. [Google Scholar] [CrossRef] [PubMed]
  121. Fuentes, A.; Yoon, S.; Kim, T.; Park, D.S. Open Set Self and Across Domain Adaptation for Tomato Disease Recognition with Deep Learning Techniques. Front. Plant Sci. 2021, 12, 758027. [Google Scholar] [CrossRef]
  122. Wang, X.; Liu, J.; Liu, G. Diseases Detection of Occlusion and Overlapping Tomato Leaves Based on Deep Learning. Front. Plant Sci. 2021, 12, 792244. [Google Scholar] [CrossRef]
  123. Fuentes, A.; Yoon, S.; Lee, M.H.; Park, D.S. Improving Accuracy of Tomato Plant Disease Diagnosis Based on Deep Learning with Explicit Control of Hidden Classes. Front. Plant Sci. 2021, 12, 682230. [Google Scholar] [CrossRef]
  124. Wang, X.; Liu, J. Tomato Anomalies Detection in Greenhouse Scenarios Based on YOLO-Dense. Front. Plant Sci. 2021, 12, 634103. [Google Scholar] [CrossRef]
  125. Tian, K.; Li, J.; Zeng, J.; Evans, A.; Zhang, L. Segmentation of tomato leaf images based on adaptive clustering number of K-means algorithm. Comput. Electron. Agric. 2019, 165, 104962. [Google Scholar] [CrossRef]
  126. Liu, J.; Wang, X. Tomato Diseases and Pests Detection Based on Improved Yolo V3 Convolutional Neural Network. Front. Plant Sci. 2020, 11, 898. [Google Scholar] [CrossRef]
  127. Qi, J.; Liu, X.; Liu, K.; Xu, F.; Guo, H.; Tian, X.; Li, M.; Bao, Z.; Li, Y. An improved YOLOv5 model based on visual attention mechanism: Application to recognition of tomato virus disease. Comput. Electron. Agric. 2022, 194, 106780. [Google Scholar] [CrossRef]
  128. Xiang, R. Image segmentation for whole tomato plant recognition at night. Comput. Electron. Agric. 2018, 154, 434–442. [Google Scholar] [CrossRef]
  129. Zhou, J.; Li, J.; Wang, C.; Wu, H.; Zhao, C.; Wang, Q. A vegetable disease recognition model for complex background based on region proposal and progressive learning. Comput. Electron. Agric. 2021, 184, 106101. [Google Scholar] [CrossRef]
  130. Liu, W.; Anguelov, D.; Erhan, D.; Szegedy, C.; Reed, S.; Fu, C.Y.; Berg, A.C. SSD: Single shot multibox detector. In Computer Vision–ECCV 2016; Lecture Notes in Computer Science; Springer International Publishing: Cham, Switzerland, 2016; pp. 21–37. [Google Scholar] [CrossRef] [Green Version]
  131. Andrearczyk, V.; Whelan, P.F. Using filter banks in Convolutional Neural Networks for texture classification. Pattern Recognit. Lett. 2016, 84, 63–69. [Google Scholar] [CrossRef] [Green Version]
  132. Singh, A.K.; Mittal, S.; Malhotra, P.; Srivastava, Y.V. Clustering Evaluation by Davies-Bouldin Index (DBI) in Cereal data using K-Means. In Proceedings of the 2020 Fourth International Conference on Computing Methodologies and Communication (ICCMC), Erode, India, 11–13 March 2020; pp. 306–310. [Google Scholar] [CrossRef]
  133. Redmon, J.; Farhadi, A. YOLOv3: An Incremental Improvement. arXiv 2018, arXiv:1804.02767. [Google Scholar]
  134. Zhang, W.; Hansen, M.F.; Smith, M.; Smith, L.; Grieve, B. Photometric stereo for three-dimensional leaf venation extraction. Comput. Ind. 2018, 98, 56–67. [Google Scholar] [CrossRef]
  135. Wu, H.; Fang, L.; Yu, Q.; Yuan, J.; Yang, C. Plant leaf identification based on shape and convolutional features. Expert Syst. Appl. 2023, 219, 119626. [Google Scholar] [CrossRef]
  136. Lin, T.Y.; Dollár, P.; Girshick, R.; He, K.; Hariharan, B.; Belongie, S. Feature pyramid networks for object detection. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA, 21–26 July 2017. [Google Scholar] [CrossRef] [Green Version]
  137. Das, M.; Deb, C.K.; Pal, R.; Marwaha, S. A Machine Learning Approach for the Non-Destructive Estimation of Leaf Area in Medicinal Orchid. Appl. Sci. 2022, 12, 477. [Google Scholar] [CrossRef]
Sustainability 15 11681 g001 550
Figure 1. Flowchart of the review process.
Figure 1. Flowchart of the review process.
Sustainability 15 11681 g001
Sustainability 15 11681 g002 550
Figure 2. Number of articles referred for review and published in each year from 2014–2023.
Figure 2. Number of articles referred for review and published in each year from 2014–2023.
Sustainability 15 11681 g002
Sustainability 15 11681 g003a 550Sustainability 15 11681 g003b 550
Figure 3. Architecture of (a) YOLOv3, (b) ResNet50.
Figure 3. Architecture of (a) YOLOv3, (b) ResNet50.
Sustainability 15 11681 g003aSustainability 15 11681 g003b
Sustainability 15 11681 g004a 550Sustainability 15 11681 g004b 550
Figure 4. Architecture of (a) VGG16, (b) AlexNet.
Figure 4. Architecture of (a) VGG16, (b) AlexNet.
Sustainability 15 11681 g004aSustainability 15 11681 g004b
Table
Table 1. Description of Research Questions.
Table 1. Description of Research Questions.
Research Question
RQ1: What are the main contributions of the current studies for tomato leaf diseases?
RQ2: What diseases are being detected for tomato crops?
RQ3: What types of datasets have been used by the contributors of the current studies?
RQ4: What are the ML and DL algorithms used to detect and classify tomato plant diseases?
RQ5: What features are being used by the ML/DL techniques to recognize and classify tomato plant diseases?
RQ6: What are the challenges and trends in tomato plant disease detection?
Table
Table 2. Summary of Common Tomato Diseases.
Table 2. Summary of Common Tomato Diseases.
CitationsDiseaseCausal AgentSymptomsSample Image
[34,35]Tomato Early BlightAlternaria solaniLeaves, stems, and fruits with circular or angular spotsSustainability 15 11681 i001
[35,36]Tomato Late BlightPhytophthora infestansLeaves and stems with large, water-soaked spots Sustainability 15 11681 i002
[35,37]Tomato Septoria Leaf SpotSeptoria lycopersiciLeaves and stems with circular, grayish-brown spotsSustainability 15 11681 i003
[35,38]Tomato Bacterial SpotXanthomonas perforansLeaves and stems with large, water-soaked spotsSustainability 15 11681 i004
[35,39]Tomato Mosaic VirusToMVMottled yellowing of leaves, stunted growth, reduced yieldsSustainability 15 11681 i005
[35,40]Tomato Yellow Leaf Curl VirusTYLCVYellowing and curling of leaves, stunted growth, reduced yieldSustainability 15 11681 i006
[35,41]Tomato Leaf MoldFolipendula fusarium (previously known as Passalora fusarium)Undersides of leaves and stems have moldy, gray-green lesions, decreased plant growth, reduced fruit productionSustainability 15 11681 i007
[35,42]Tomato Two Spotted Spider MiteTetranychus urticaeYellowing and stippling of leaves, reduced plant growth and yieldsSustainability 15 11681 i008
[35,43]Tomato Target SpotCorynespora cassiicolaLeaves, stems, and fruit with circular, dark-brown to black lesions with concentric ringsSustainability 15 11681 i009
Table
Table 3. Datasets used in various articles reviewed.
Table 3. Datasets used in various articles reviewed.
CitationsDatasetType of Dataset
[7,19,21,22,27,31,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100]PlantVillagePublic
[101,102]AIChallengerPublic
[17,18]Own datasetPublic
[103]Tomato Leaf Disease DetectionPublic
[104]Dataset of Tomato LeavesPublic
[105]PlantVillage + Plant Disease Severity Public
[106]PlantVillage + New Plant DiseasesPublic
[107]PlantVillage + AIChallengerPublic
[33]PlantVillage + AIChallenger + PlantDocPublic
[32,108]PlantVillage + PlantDocPublic
[109,110,111,112,113,114,115,116,117]PlantVillage + own datasetPublic + Private
[8,23,24,118,119,120,121,122,123,124,125,126,127]Own datasetPrivate
Table
Table 4. Summary of Algorithms with Highest Accuracies.
Table 4. Summary of Algorithms with Highest Accuracies.
ReferenceAlgorithmAccuracy (%)
[110]EfficientNetB3 + LR 100
EfficientNetB3 + kNN 100
EfficientNetB3 + RF100
EfficientNetB3 + SGB100
EfficientNetB3 + ADB100
[61]DCGAN-PILAE100
[76]DENN + Transfer Learning100
[72]EfficientNetB399.997
[72]EfficientNetB599.997
[77]Custom CenterNet + DenseNet7799.982
[51]ICRMBO (Improved Crossover-based Monarch Butterfly Optimization) + VGG1699.98
[78]Faster R-CNN + ResNet3499.97
[51]ICRMBO + InceptionV399.94
[33]ICVT99.94
Table
Table 5. Summary of Features and their Techniques.
Table 5. Summary of Features and their Techniques.
ReferenceFeaturesTechniques
[23,24,26,48,49,56,58,62,63,69,85,97,122,127]TextureHaralick Features, LBP, GLCM, Law’s Mask, Spatial-wise Feature Extraction
ColorColor Histograms, Hue Saturation Value, Color Coherence Vector, Spatial-wise Feature Extraction
ShapeHu Moments, Spatial-wise Feature Extraction
[27,52,83,94,102,106,113]TextureGLCM, Local Binary Pattern, Speeded Up Robust Features, Otsu’s Algorithm, Information Gain, Histogram of Oriented Gradients, Entropy
ColorColor Moment, Color Histogram, Color Histogram Combinations–LBP, GLCM, LBP + GLCM, COLOR, COLOR + LBP, COLOR + GLCM, ALL, Combined Genetic Algorithm, Correlation-based Feature Selection, K-means Clustering
[54]TextureScale Invariant Feature Transform (SIFT), GLCM
EdgeHarris Corner Detector
[55]TextureCNN
ColorCNN
MorphologicalCRNN
SequentialCRNN
Gradient-basedCRNN
[91,103,125]ColorHyperspectral Imaging, L*a*b*, Hue Saturation Value, YCbCr, Luminance In-phase Quadrature (YIQ), Hue Saturation Intensity, CIELAB (International Commission on Illumination (CIE))
[73]Coefficients of LL Sub-band at Level 3 (Haar Wavelet)2D–Discrete Wavelet Transform
[19]VeinMorphological Opening
TextureGLCM
Haar-like FeaturesCanny Edge Detection
EdgeGeometric Features
ShapeFourier Descriptors of Polar Fourier Transform
[88]Sub band imagesAdaptive Analytic Wavelet Transform
[84,89]TextureSIFT, GLCM
[116]TextureCNN
ShapeHue Saturation
[108]TextureT-CNN
ShapeT-CNN
VeinT-CNN
[117]Minor LesionLocation-wise Soft Attention Mechanism
[32]ColorHierarchical Bilinear Pooling
ShapeHierarchical Bilinear Pooling
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Sundararaman, B.; Jagdev, S.; Khatri, N. Transformative Role of Artificial Intelligence in Advancing Sustainable Tomato (Solanum lycopersicum) Disease Management for Global Food Security: A Comprehensive Review. Sustainability 2023, 15, 11681. https://doi.org/10.3390/su151511681

AMA Style

Sundararaman B, Jagdev S, Khatri N. Transformative Role of Artificial Intelligence in Advancing Sustainable Tomato (Solanum lycopersicum) Disease Management for Global Food Security: A Comprehensive Review. Sustainability. 2023; 15(15):11681. https://doi.org/10.3390/su151511681

Chicago/Turabian Style

Sundararaman, Bharathwaaj, Siddhant Jagdev, and Narendra Khatri. 2023. "Transformative Role of Artificial Intelligence in Advancing Sustainable Tomato (Solanum lycopersicum) Disease Management for Global Food Security: A Comprehensive Review" Sustainability 15, no. 15: 11681. https://doi.org/10.3390/su151511681

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop