Journal of Agricultural Machinery
with the collaboration of Iranian Society of Mechanical Engineers (ISME)

Machine Learning for Detection of Pests in Tomato: A Review

Document Type : Review Article- En

Authors

M. Keerthivasan 1
S. Kokilavani 1
M. Shanthi 2
Ga. Dheebakaran 1
R. Pangayar Selvi 3
M. Murugan 4
T. Elaiyabharathi 4
P. S. Shanmugam 4
M. Selva Kumar 1

1 Agro Climate Research Centre, Tamil Nadu Agricultural University, Coimbatore, Tamil Nadu, India

2 Centre for Plant Protection Studies, Tamil Nadu Agricultural University, Coimbatore, Tamil Nadu, India

3 Department of Physical Science and Information Technology, Tamil Nadu Agricultural University, Coimbatore, Tamil Nadu, India

4 Department of Agricultural Entomology, Tamil Nadu Agricultural University, Coimbatore, Tamil Nadu, India

https://doi.org/10.22067/jam.2025.90762.1314
Abstract
Influence of a single atmospheric component or meteorological variable on the host, pathogen, or their interaction in controlled environments has accounted for the majority of climate change’s impact on plant pests and diseases. Climate change can lead to alterations in the stages and rates of growth of pests and diseases, host resistance, and the physiology of host-pathogen or host-pest interactions, which can cause substantial harm and reduce tomato crop yields. Different approaches have been ineffective in the accuracy of pest and disease forewarning in past years. The remarkable progress in Deep Convolutional Neural Networks (DCNNs) is revolutionizing the early detection of pests and diseases in crops. By analysing vast amounts of present and historical climate data, alongside their expertise in object identification and image categorization, these AI models can predict outbreaks with impressive accuracy. However, understanding the specific microclimate suitable for each pest and disease is crucial for truly effective intervention. Combining these two elements creates a powerful, targeted approach to preserving crops. A forewarning system can help to reduce the use of pesticides, thereby reducing the cost of production and environmental pollution. Proper cloud servers and IoT-based sensor networks should be used for a better forewarning of pests and diseases in future circumstances.

Keywords

CNN
Machine learning
Microclimate
Pest forewarning
Tomato
Subjects

Introduction

Climate change has been shown to have a substantial influence on the behaviour of pests and diseases in agricultural systems (Bebberet al., 2013). Alterations in temperature and precipitation patterns substantially affect the behaviour of pests and disease vectors, creating unexpected challenges for farmers and agricultural productivity (Heebet al., 2019). Climate change may also have an impact on pest and disease vector life cycles and behaviour. Warmer winters might lower insect death rates, enabling their populations to increase more quickly in subsequent seasons. Similarly, changed precipitation patterns may generate circumstances that promote the reproduction and spread of disease-carrying organisms (Trebicki and Finlay, 2019). Besides, invasive species pose one of the biggest dangers to society, the economy, and global biodiversity (Earlyet al., 2016; Sarukhánet al., 2005). The inadvertent long-distance spread of invasive pests and diseases into areas outside of their natural distribution ranges has been greatly stimulated by climate change (Battisti and Larsson, 2015; Musolin, 2007). The establishment rate of invasive species has roughly quadrupled in the last 30-40 years in the European continent alone (Roqueset al., 2016) and the number of invasive forest diseases has escalated dramatically over the last 200 years (Santiniet al., 2013). Worldwide, it is acknowledged that the primary means of introducing invasive pests and diseases into the agricultural ecosystem is through the trade of planting materials (Brasier, 2008; Keniset al., 2007; Liebholdet al., 2012; Santiniet al., 2013; Santiniet al., 2018). However, the equation is further complicated by the factor of climate change, which can act as a catalyst, altering the delicate interaction between hosts and pests within crop systems.

Pest forecasting is an important part of pest management. Pest forecasting refers to accurately predicting pest outbreaks using relevant data, a process vital for protecting the agri-food supply chain and the environment (Liet al., 2022). With proper forecasting, farmers can prepare and manage impending pest outbreaks, minimize unnecessary pesticide use, and reduce environmental pressures (Isardet al., 2015; Machekanoet al., 2019). Within the realm of integrated pest management, a multifaceted approach is crucial for effective pest control. Machine learning has revolutionized the way pest and disease forecasting in agriculture has leveraged large datasets to develop predictive models (Ahmed, 2018). By analysing vast amounts of data, machine learning algorithms can identify patterns and make predictions with a high degree of accuracy (Pak and Kim, 2017). While machine learning (ML) offers powerful tools for rapid and accurate pest identification, understanding both weather and climate suitability is equally important for developing comprehensive control strategies. Integrating ML-based identification with weather and climate suitability modeling creates a robust and targeted approach to pest management, fostering sustainable agricultural practices.

Tomato is the most significant vegetable crop in the world economy and its output has grown significantly over time (Huet al., 2023). Lycopersicon esculentum thrives in a warm, sun-drenched environment with well-drained soil. The optimal temperature range for tomato growth lies between 20°C and 31°C, with nighttime temperatures of 13°C to 18°C contributing to enhanced flavour and colour development. Moderate rainfall is sufficient for tomato cultivation, while excessive humidity can foster the proliferation of fungal diseases (Yanget al., 2019). Generally, tomatoes will be affected by high temperature and water stress (Hernandez-Espinoza and Barrios-Masias, 2020). Thus, an optimum microclimate is essential for effective fruit production and crop output. In addition to these, tomato plants are vulnerable to over 200 pests and diseases caused by pathogenic fungi, bacteria, viruses, and nematodes. Every pest and disease require a unique microclimate condition for their growth and development. Warm, humid circumstances often favour fungal diseases such as late blight in tomatoes, while cool nights and high humidity encourage the growth of Septoria leaf spots (Singhet al., 2018). Integrated pest management (IPM) requires a robust pest and disease surveillance system to deploy timely plant protection measures when needed to lower cultivation costs and prevent ecosystem contamination.

Recent integrated pest management studies indicate that weather variations significantly influence pest and disease outbreaks (Dhawan, 2016; Fuenteset al., 2017; Saeedet al., 2018). An essential component of pest control is recognizing the tomato crop’s pest complex and its correlation with meteorological variables (Rawat, Karnatak, and Srivastava, 2020). When the ideal meteorological condition for a pest invasion is understood, it can facilitate pest detection (Alamet al., 2016). For efficient management of pests and diseases in modern agriculture, microclimate-based forewarning systems are essential. Tomato crop productivity and quality may be greatly increased with early detection and intervention, all while using fewer pesticides. Thus, understanding the relationship between host and pest, along with predicted weather patterns, allows for the prevention of pest and disease outbreaks (Balikaiet al., 2021).

Thus, the current article explores the past literature investigating machine learning applications in pest management and suitable microclimates for different pests and diseases affecting tomatoes.

Methodology

The search utilized databases including Google Scholar and Scopus. The search string used to collect information was outlined below:

(“Pest forewarning AND Tomato”, “Regression models” AND “Pest forewarning”, “Logistic models” AND “Pest forewarning”, “ARIMA model” AND “Pest forewarning”, “Machine learning” AND “Pest forewarning” AND “Tomato”, “Pests” AND “Tomato AND “Deep learning”, “Diseases” AND “Tomato” AND “Deep learning”)

To be included in the current review, articles needed to provide quantitative results about at least one aspect of pest forewarning models, neural network or influence of pest and diseases of Tomato based on micro climate.

Fig. 1. Conceptualized framework for pest prediction and identification

Fig. 2. Identification of studies via databases

Pest forewarning/detection

Advances in image processing and intelligent monitoring technologies have emerged as promising tools for early pest detection, enabling timely intervention and reducing the need for excessive pesticide application (Nagar and Sharma, 2020). These techniques leverage computer vision and machine learning algorithms to automatically identify pests, reducing the reliance on human experts and streamlining the pest management process (Ngugiet al., 2021). By providing farmers with accurate and real-time information about pest infestations, these technologies can help them make informed decisions, leading to more sustainable and eco-friendly agricultural practices. Moreover, the integration of these systems with the Internet of Things (IoT) and decision support systems can further enhance the effectiveness of pest management, enabling coordinated efforts across larger geographic regions (Limaet al., 2020).

Logistic and regression models

Previously, ordinal logistic models were used to forecast the pest/disease’s occurrence. In cases where the data was quantitative, it was converted into dichotomous using threshold values. The model had the following form:

P(Y=1)==11+expexp(-L)andL=∑βiXi(1)

where Xi denotes the weather variables/weather indices. P< 0.5 indicates a low likelihood of an epidemic occurring, whereas P> 0.5 suggests a higher likelihood. The function L was constructed using several combinations of weather variables, including maximum and minimum temperatures, relative humidity (morning and evening), and mean relative humidity, as well as interactions. The combination that most accurately predicted the observed data was identified (Srivastavaet al., 2015). Later, stepwise regression models were used to forecast various aspects of rice, mustard, pigeon pea, sugarcane, groundnut, and cotton pests and diseases at various locations, including maximum pest population/disease severity, time of first appearance, time of maximum pest population/disease severity, and weekly pest population/disease severity. In this sort of model, two indices have been generated for each weather variable, one as a total of weather variable values in different weeks and the other as a weighted total, with weights representing correlation coefficients between the variable to forecast and the weather variable in those weeks. The first index represents the total amount of weather variable over the period under examination, while the second addresses the distribution of weather variables, with a focus on their importance in different weeks in relation to the variable to be anticipated. Similarly, indices were calculated for joint effects using weather variable products (two at a time). The model’s form is:

Y=a0+∑i≠1p∑j=01aijZij+∑i≠1p∑j=01bii, jZii, j+e(2)

where

Zij=∑w=n1n2riwjXiwandZii, j=∑w=n1n2riwjXiwXi, w

Y: variable to forecast

X: value of ith weather variable in wth week

riw: correlation coefficient between Y and ith weather variable in wth week

rii , w: correlation coefficient between Y and product of Xi and Xi , in wth week

P: number of weather variables considered

n1: initial week for which weather data were included in the model

n2: final week for which weather data were included in the model

In some cases, previous disease incidence/pest population (or their indices) and/or the previous year’s last population have also been included in the model. Stepwise regression technique was used for selecting important variables to be included in the model. This approach allows for credible warnings at least one week in advance (Chattopadhyayet al., 2005a; Chattopadhyayet al., 2005b; Desaiet al., 2004; Vishwa Dharet al., 2007).

When data is provided for a few years (5-7 years) at varied time intervals (weekly), it is insufficient for building standard models. In this case, the deviation method can be used. It has been considered that the pest population/disease severity at any given moment is determined by the pest’s natural life cycle and the meteorological conditions. In order to discover the natural pattern, data from various intervals are averaged across time, and an appropriate model can be established. A model may be fitted with deviations from natural patterns as the dependent variable and weather as the independent variable. Mehtaet al. (2001) demonstrated this methodology for weekly fruit fly populations in mango at Rehman Khera Farm, Central Institute for Subtropical Horticulture, Lucknow, India.

On the other hand, if historical data is unavailable and only 10-12 data points exist between the time of first appearance of disease/pest and maximum disease severity/pest population, a forecast of maximum disease severity/pest population can be obtained from current season data using a within-year growth model. The technique entails fitting a suitable model to the pattern of disease development/pest population using partial crop season data and forecasting the maximum value based on that model. This technique was used to forecast the percent disease severity (PDS) of Alternaria Blight in the Varuna mustard variety at Kumarganj in 1999-2000 for various sowing dates. The model was,

(3) Yt=AexpexpBte(3)

where, t: weeks after sowing, Yt: percent disease severity at week t, and A and B: model parameters. Using this model, reliable forecasts could be obtained two weeks in advance (Mehtaet al., 2005).

ARIMA model (Autoregressive Integrated Moving Average)

ARIMA models utilize historical data to identify patterns and trends, making them an effective tool for forecasting in agricultural settings. By analysing factors such as weather patterns, crop health, and pest populations, ARIMA models have the benefit of being able to capture and account for complex temporal patterns and seasonal fluctuations. This makes them ideal for forecasting pest and disease dynamics, which are naturally impacted by a variety of environmental and biological variables (Collier, 2017). The ARIMA (p, d, q) model has three parameters. The autoregressive parameter, denoted by parameter p, examines the connection between a variable and its previous occurrences. Through analysing historical data, the autoregressive parameters can be derived to predict future instances of pest and disease occurrences. The second parameter is differentiation (d), and the number of lag forecast mistakes is represented by the running mean value or parameter q. Choosing the right lags is a vital stage in constructing a reliable ARIMA model for forecasting pest and disease outbreaks (Setiyowatiet al., 2015). Lags indicate the number of past data points considered for predicting future occurrences. The ARIMA model was created using an autocorrelation plot (ACP) on stationary time-series data (Lee and Liu, 2014). The moving average parameter was obtained using the value of the partial correlation coefficient. The moving average segment within the ARIMA model depicts the association between a current observation and the residual error obtained from applying a moving average model to past observations (Mahapatra and Dash, 2020). The created ARIMA model was evaluated by comparing observed data from the farm field with data that the model is anticipated. According to the findings above, ARIMA (1, 0, 2) is the best model to forecast the incidence of pests and diseases based on microclimatic data.

Machine learning

Machine learning has shown to be a great revolution in agriculture, transforming many processes to enhance efficiency, production, and sustainability (Chlingaryanet al., 2018). From agricultural yield prediction to disease detection, machine-learning approaches have shown enormous promise in revolutionizing agriculture (Priya and Ramesh, 2020). Aerial colour and color-infrared photography have long been used to monitor crop development; however, these technologies are being extensively researched for analysing spatial variability within the field. High-resolution imagery’s finer features enable a closer study of crops, allowing for earlier and more accurate diagnosis of stress symptoms, pest infestations, and diseases. This can be critical for making timely interventions and reducing yield losses (Huntet al., 2004). Overall, the use of machine learning in agriculture has resulted in substantial advances and has immense potential for the future (Bestelmeyeret al., 2020).

Precision farming is one of machine learning’s most successful applications in agriculture, which may give farmers useful insights into improving their irrigation, fertilization, and pest management techniques by analysing data from sensors, satellite photos, and weather predictions (Priya and Ramesh, 2020).

One of the major accomplishments of machine learning models is the detection of agricultural diseases and insect infestations, allowing for early intervention and more effective resource allocation (Maduranga and Abeysekera, 2020). In any crop, pest and disease prediction is a crucial component for the process of managing pests and diseases. Thus, it is necessary to understand the life cycle of the pests and schedule management practices to align with the stages of the pest and pathogen life cycle (Collier, 2017). Machine learning algorithms are becoming more useful in agricultural pest and disease prediction. These algorithms are capable of analysing enormous amounts of data and detecting trends that may indicate pest or disease breakouts (Javaidet al., 2023). They can aid in the prediction of possible pest and disease infestations by collecting data on weather patterns, crop health, insect populations, and other relevant aspects (Singhet al., 2018). When compared to traditional methods, these systems can greatly increase crop output and quality with reduced pesticide usage.

A study by Bhatiaet al. (2020) investigates the application of machine learning for identifying tomato pests from images using three classifiers: SVM, k-NN, and DT. Texture features such as GLCM, LBP, HOG, and SURF were utilized, with SVM combined with LBP achieving the best accuracy of 81.02%. The study emphasizes the importance of early pest detection in enhancing tomato crop quality and yield. It showcases the potential of integrating image processing with machine learning to advance agricultural practices.

Pest forewarning and detection utilize a diverse toolbox of methods, and some of the previous studies related to the use of machine learning for pest and disease detection are compiled in Table 1.

Crop Disease/insect Algorithm Data set Observation Reference
Mango Thrips Random forest Temperature and humidity Low error estimates (Jawade et al., 2020)
Rice Blast Artificial neural network (ANN) Temperature and humidity Average accuracy of 72% (Jia et al., 2020)
Mango Powdery mildew Rough sets, linear regression, rough sets based decision, decision tree Temperature and humidity Average accuracy of 75% (Rajni et al., 2009)
Rice Blast Long-term memory network (LSTM) Temperature and humidity Maximum accuracy of 79.4% ([Kim et al., 2020)
Coffee Rust Fuzzy Temperature and humidity Fuzzy models outperformed classical models in terms of error rates. (Cintra et al., 2011)
Pomegranate General pomegranate diseases Hidden Markov chain Temperature, humidity and wind speed Accuracy of 80.7% (Pawara et al., 2018)
Cherry A general disease of cherry Discriminant Analysis Temperature, humidity, rainfall and wind speed Maximum prediction accuracy of 93.6% (Ilic et al., 2018)
Tomato Powdery mildew Hybrid of support vector machine (SVM), LR Temperature, Humidity, leaf wetness, wind speed, global radiations Conducive classes, not conducive classes (Bhatia et al., 2020)
Rice Blast ANN, SVM Evaporation, maximum temperature, minimum temperature, Rainfall, solar radiation, wind speed and humidity Disease occurrence, disease severity (Malicdem and Fernandez, 2015)
Potato Late blight ELM, SVM maximum temperature, minimum temperature, maximum humidity, minimum humidity, Rainfall, Number of rainy days Class 1: <3%, class 2: 4-10%, class 3: 11-30%, class 4: 31-60%, class 5: >60% (Singh et al., 2019)
Orange Black spot, greening, melonasa, greasy spot, scab, Alternaria brown spot, canker ACC Image of Orange, temperature, rainfall, humidity Disease name (Kaur and Kaur, 2018)
Olive, Grape Downy mildew, powdery mildew, peacock spot, anthracnose ACC Temperature, humidity, accumulated heat, degree days Disease occurrence (Alves et al., 2018)
Cucumber Downy mildew, leaf spot, anthracnose Wavelet transformation, SVM 13 variables divided into soil data, weather data, disease data Disease occurrence (%) (Junjing et al., 2019)
Table 1. Use of machine learning techniques to predict agricultural pests and diseases

Deep learning in object detection

ANNs have emerged as effective tools for solving complicated issues, such as pest prediction in agriculture (Wanget al., 2022). Because of its ability to model and analyze deep interactions within big datasets, ANNs are commonly used for crop pest risk prediction. Research has shown that ANNs have good prediction accuracy in anticipating insect populations and assessing their hazards in a variety of crops, including rice. Furthermore, ANNs have been employed in intelligent agent-based prediction systems for pest detection and alert mechanisms, leveraging technologies such as acoustic methods and video processing techniques to allow early pest finding and classification. In pest management, object detection can be used to identify and monitor numerous pests, providing useful information for decreasing their populations and limiting crop loss. Deep learning, a subset of machine learning, has shown significant potential in object detection applications due to its capacity to understand complicated patterns and characteristics from vast datasets (Corraleset al., 2015). Digital image processing has seen improvements through deep learning, which is significantly better than conventional techniques.

Early pest and disease detection and prediction is made possible by this technology, allowing for prompt intervention and control actions (de Souzaet al., 2017). Deep learning can be also used in the detection of pests and diseases based on weather data (Panditet al., 2022).

Convolutional neural network

Convolutional neural networks are leading the way in computer vision problems (Szegedyet al., 2015). Unlike traditional approaches to training classifiers with hand-designed feature extraction, CNNs learn feature hierarchy from pixels to classifier, and train layers simultaneously. Because of the intricacy of the model, CNNs can take weeks to fully train; thus, transfer learning is used to shorten model training by taking a fully trained model for a set of classes and retraining the current weights for additional classes.

One of the primary advantages of CNNs is their capacity to analyze and interpret visual data, which makes them ideal for tasks like plant disease assessment, crop monitoring, and yield prediction (Prashar and Sangal, 2022). CNNs may be taught to detect patterns and abnormalities in photographs of crops, soil, and agricultural landscapes (Zhaoet al., 2020). This enables the early diagnosis of pest or disease infestations, as well as prompt action to avoid widespread damage (Aiet al., 2020). Another distinguishing aspect of CNNs is their capacity to process vast amounts of visual data efficiently (Boulentet al., 2019).

Several studies on Machine Learning (ML) techniques used in the agricultural sector focus on the tasks of classification, detection, and prediction of diseases and pests, with a focus on tomato crops (Domingueset al., 2022). In recent years, there has been an increased interest in employing convolutional neural networks to anticipate and detect possible pest and disease outbreaks in tomato crops (Brahmiet al., 2024). To train the Convolutional Neural Network, a dataset was collected from various tomato farms in different regions (Verma and Zhang, 2018). Daily microclimate measurements such as temperature, humidity, and sunlight exposure, along with information on the presence of pests and diseases in the tomato plants were required (Gurleet al., 2019). The data collection process involved collaborating with multiple farms to ensure a diverse and representative dataset for training the model. Additionally, advanced sensors were strategically placed across the farms to capture real-time microclimate data. This ensured that the dataset not only had a wide geographical representation but also captured the dynamic nature of microclimate within each region (Sladojevicet al., 2016). CNNs have shown considerable potential in image recognition and classification tasks, making them an ideal choice for detecting pest and disease indicators in tomato plants (Wiesner-Hankset al., 2018).

The collected data will then be pre-processed to remove any outliers or inconsistencies before being used for training the Convolutional Neural Network (de Souzaet al., 2017). Data pre-processing also includes data augmentation techniques such as random contrast, flip, zoom, and rotation to enhance the diversity of the training dataset and prevent overfitting (Shorten and Khoshgoftaar, 2019). A study by Kattenbornet al. (2021) highlights that batch normalization standardivolzes activation function outputs to have a zero mean and unit variance, preventing imbalance caused by extreme activation values. This technique simplifies the gradient descent optimization process, enables the use of larger learning rates, and accelerates network convergence.

The training process involved splitting the dataset into training, validation, and testing sets to evaluate the performance of the model. Various hyperparameters were tuned, and different architectures were explored to optimize the CNN for accurate pest and disease forewarning (López-Moraleset al., 2008). The CNN was trained to automatically extract features from the raw microclimate data and correlate them with the presence of pests and diseases in the tomato plants (Brahimiet al., 2017). Training a CNN on a varied collection of tomato plant photos, including healthy and diseased plants, may lead to a powerful forewarning system (Kamilaris and Prenafeta-Boldú, 2018). To accomplish this goal, the researchers first compiled a broad collection of tomato plant photos, including those displaying healthy plants as well as those damaged by numerous pests and diseases (Niharet al., 2021).

Fig. 3. Architecture of the CNN used for detecting pinworm damage; the input consisted of leaves infected with pinworm, non-linearity includes the activation layer and convolutional layer filters the input images and reduces the size of the dataset

This dataset will be critical in training the convolutional neural network to recognize and categorize various signs and manifestations of pests and diseases in tomato (Paymodeet al., 2021). Once the dataset is produced, it will begin training the CNN, using approaches like data augmentation to improve the model’s capacity to generalize and predict on previously unknown data (Jiaet al., 2020). The model will be validated and fine-tuned to guarantee that it is successful in detecting possible pest and disease outbreaks. Through the training and evaluation process, the developed CNN models showed satisfactory performance in detecting nine different tomato diseases and pests. (David, 2023).

Fig. 4. Pest and disease forewarning using convolutional neural network

The CNN models achieved a high accuracy rate of 99.18% in classifying the presence of pests and diseases based on the microclimate data. This high accuracy rate outperformed previous shallow models and demonstrated the effectiveness of using Convolutional Neural Networks for pest and disease forewarning in tomato plants (Emeboet al., 2019). The results of this study highlight the potential of using deep learning and Internet of Things technologies, specifically convolutional neural networks, for accurate and efficient pest and disease forewarning in tomato plants (Gonzalez-Huitronet al., 2021). This approach provides farmers with a practical tool to remotely monitor the health of their tomato plants, reducing effort and enabling timely interventions (Agarwalet al., 2020). Once the model was trained, it underwent rigorous testing and validation to assess its ability to accurately predict pest and disease occurrences based on the microclimate data (Prajwalaet al., 2018). The results of the training and validation process were crucial in determining the effectiveness and reliability of the CNN for forewarning pest and disease outbreaks in tomato plants (de Souzaet al., 2017). Farmers will get this pest and disease warning information and advice on proper crop management techniques via electronic media, such as the Internet, short messaging service (SMS) and other digital platforms (Hughes and Salathé, 2015).

Recurrent-CNN

Girshicket al. (2014) proposed R-CNN as a method for efficient object localization in object identification. R-CNN uses the selective search algorithm. The initial step in deploying R-CNN for agricultural crops is to gather and preprocess the dataset. This involves acquiring photos of numerous crops, including wheat, rice, maize, and soybeans, from diverse agricultural areas (Qiet al., 2023). The collection should also contain photographs of prevalent pests and diseases that attack these crops (Barbedo, 2020). Once the photos have been acquired, they must be pre-processed to ensure that they are in the correct format for training the R-CNN model. This might include scaling the photographs, normalizing pixel values, and labelling them with bounding boxes to show the location of crops, pests, and diseases. Following the preparation of the dataset, the R-CNN model is trained. A convolutional neural network is used to extract features from the pictures, followed by a region proposal network to provide plausible bounding boxes for objects in the images. The model is then trained to categorize the items that fall inside the suggested boundaries (Patel and Bhatt, 2021). Once trained, the model must be assessed against a different testing dataset to determine its accuracy and performance. Following assessment, the model may be used to detect and analyze crops, pests, and diseases in agricultural settings (Zhenget al., 2019).

Fast R-CNN

Fast R-CNN was created to address the computational inefficiencies of R-CNN. Unlike R-CNN, which processes each region proposal separately via CNN, fast R-CNN uses the whole picture and its region suggestions as input in a single forward pass through the CNN architecture. This considerably minimizes the processing overhead. Furthermore, fast R-CNN combines numerous architectural components, such as ConvNet, RoI pooling, and classification layers, into an organised and efficient framework. One of the primary benefits of adopting fast R-CNN in agriculture is its capacity to reliably identify regions that need attention, allowing for focused treatments while minimizing total input consumption such as pesticides and herbicides (Sykeset al., 2023). One of the primary benefits of adopting Fast R-CNN in agriculture is its ability to recognize items of interest in pictures quickly and accurately (Halsteadet al., 2018). This allows farmers to immediately detect and rectify potential issues in their fields, resulting in better crop management and higher yields (Qianget al., 2020).

Faster R-CNN

Faster R-CNN, proposed by Renet al. (2015), overcomes the bottleneck of selective search in its predecessor by including a region proposal network. After running the picture through a backbone network, it generates convolutional feature maps. The area proposal network then uses these feature maps to create anchors, which represent the centres of sliding windows of various sizes and scales. These anchors are then processed by the classification layer, which determines object functioning, and the regression layer, which localizes the bounding boxes. This method substitutes selective search, considerably increasing object detection efficiency by including region proposal creation into the network design.

Recently, a smartphone application called "PESTPREDICT" which uses weather-based pest warnings as part of integrated pest management for crop protection was developed. This program helps farmers, agricultural extension workers, and researchers obtain location-specific predictions of desired pest or disease for target crops so that they may be effectively managed (Szegedyet al., 2015). Plantix is an influential software that is gaining popularity among farmers and gardeners. Its key characteristics include plant disease diagnosis, pest detection, and gardening recommendations (Siddiquaet al., 2022). The app uses picture recognition technology to identify plant pests and diseases, giving users precise information and treatment recommendations (Petrellis, 2017). In addition to its diagnostic capabilities, Plantix functions as an information platform for plant enthusiasts, providing articles, videos, and other materials to assist users in taking care of their plants. Plantix has become an indispensable tool for anybody who wants to keep their plants healthy and developing (Mrishoet al., 2020). In this study by Linet al. (2020), Faster Region-Convolutional Neural Networks (Faster R-CNN) and Mask R-CNN were utilized to create a knowledge-based system capable of automatically identifying plant pests and diseases. The Faster R-CNN achieved a regional recognition accuracy of 89%, while the Mask R-CNN demonstrated an area recognition accuracy of 81%. This research successfully developed a system for pest and disease identification.

YOLO (You Only Look Once)

The YOLO approach to object detection revolutionizes the process by enabling simultaneous identification and localization of objects in an image at a single glance. Unlike traditional multi-step methods, YOLO redefines object detection as a regression problem, directly predicting spatially separated bounding boxes and their corresponding class probabilities. This is achieved through a single neural network evaluation, making the process efficient and streamlined (Du, 2018). A specialized tomato-picking robot is designed exclusively for use in facility agriculture environments. A deep learning-based recognition technique is implemented. The process identifies the positions of peduncles by connecting bounding boxes and acquires depth data using a depth camera. It then simulates manipulator trajectory planning based on spatial coordinates, overcoming challenges like lighting variations and obstructions. YOLO outperforms the SSD algorithm with higher accuracy and confidence in tomato shape detection. Future studies could adapt this approach to other crops like eggplants, cucumbers, and oranges with contrasting features (Zhaoxinet al., 2022).

Boons

Large data sets can be analyzed by machine learning algorithms, which can spot patterns and trends that are invisible to the human eye (Gauriauet al., 2024). This makes it possible for agriculture specialists to more precisely forecast and warn of impending outbreaks of pests and diseases (Ifftet al., 2018). Machine learning models can provide early warnings by utilizing environmental conditions, historical data, and crop health indicators. This enables farmers to take proactive actions to reduce the impact of pests and diseases on their crops (Veeragandham and Santhi, 2020). Farmers can save a large amount of money by utilizing machine learning to forecast pests and diseases. Farmers can decrease the needless use of chemical pesticides and fertilizers, resulting in lower costs and more sustainable agricultural practices, by precisely anticipating and diagnosing potential dangers to their crops (Bestelmeyeret al., 2020). Farmers can also reduce crop losses and guarantee a more consistent and dependable production by using preventative measures based on machine learning insights, which will further contribute to agriculture’s long-term sustainability (Rehmanet al., 2019). Based on the unique requirements and conditions of each farmer’s crop, machine learning algorithms can offer customized recommendations and decision support (Ahmedet al., 2024). These algorithms can adjust and improve their recommendations by evaluating real-time data and continuously learning from new information, accounting for fluctuations in crop types, weather, and soil conditions (Sudduthet al., 2020). With this degree of individualized support, farmers are better equipped to manage pests and diseases, and make well-informed decisions that improve crop output and health (Benoset al., 2021).

Conclusion

Overall, the implementation of microclimate-based pest and disease warning systems in tomato agriculture has the potential to transform the way farmers manage their crops. This technique enables farmers to make educated choices by using real-time data and advanced analytics, resulting in more effective pest and disease management strategies. It minimizes pesticide use, and also promotes sustainable agriculture practices and crop quality. More research and technical improvements will improve the accuracy and reliability of microclimate-based forewarning systems. This will allow for greater adoption across areas and crop kinds, eventually improving agricultural production and environmental sustainability. As the agricultural industry evolves, adopting novel solutions such as microclimate-based forewarning will be critical in tackling pest and disease concerns, resulting in a more resilient and productive agricultural sector. In addition, incorporating microclimate-based pest and disease forewarning into tomato growing might provide farmers with economic advantages. Farmers may be able to boost their total production and revenues by lowering their reliance on pesticides and crop loss. This, in turn, may help ensure the long-term profitability of agricultural enterprises and the livelihoods of rural communities. Furthermore, as technology advances and these systems become more accessible and user-friendly, it is expected that small-scale farmers will be able to utilize the advantages of microclimate-based forewarning, democratizing access to effective pest and disease control tools.

Acknowledgment

The financial support received from World Vegetable Centre, Taiwan and TNAU IPM project (Project No: WVC/CPPS/CBE/2022/R001) is gratefully acknowledged.

Funding

The manuscript was funded by World Vegetable Centre, Taiwan and TNAU IPM project (Project No: WVC/CPPS/CBE/2022/R001).

Conflict of Interest: The authors declare no competing interests.

Author Contributions

M. Keerthivasan: Writing– original draft, Conceptualization, Methodology, Visualization

S. Kokilavani: Conceptualization, Methodology, Review and editing

M. Shanthi: Technical advice, Review and editing

Ga. Dheebakaran: Review and editing

R. Pangayar Selvi: Supervision

M. Murugan: Technical advice

T. Elaiyabharathi: Validation

P. S. Shanmugam: Supervision

M. Selva Kumar: Review and editing

References

  1. Agarwal, M., Gupta, S. K., and Biswas, K. (2020). Development of Efficient CNN model for Tomato crop disease identification. Sustainable Computing: Informatics and Systems, 28, 100407.DOI
  2. Ahmed, F. (2018). An IoT-big data based machine learning technique for forecasting water requirement in irrigation field. In Research and Practical Issues of Enterprise Information Systems: 11th IFIP WG 8.9 Working Conference, CONFENIS 2017, Shanghai, China, October 18-20, 2017, Revised Selected Papers 11 (pp. 67-77). Springer International Publishing.DOI
  3. Ahmed, S., Basu, N., Nicholson, C. E., Rutter, S. R., Marshall, J. R., Perry, J. J., and Dean, J. R. (2024). Use of machine learning for monitoring the growth stages of an agricultural crop. Sustainable Food Technology, 2(1), 104-125.DOI
  4. Ai, Y., Sun, C., Tie, J., and Cai, X. (2020). Research on recognition model of crop diseases and insect pests based on deep learning in harsh environments. IEEE Access, 8, 171686-171693.DOI
  5. Alam, M. Z., Haque, M. M., Islam, M. S., Hossain, E., Hasan, S. B., Hasan, S. B., and Hossain, M. S. (2016). Comparative study of integrated pest management and farmers practices on sustainable environment in the rice ecosystem. International Journal of Zoology, 2016.DOI
  6. Alves, L., Silva, R. R., and Bernardino, J. (2018). System to predict diseases in vineyards and olive groves using data mining and geolocation. In Proceedings of the 13th International Conference on Software Technologies (ICSOFT 2018) (pp. 713–721). SCITEPRESS.DOI
  7. Balikai, R., Venkatesh, H., and Sagar, D. (2021). Development of models to predict insect pest populations-an eco-friendly tactic for pest management. Journal of Farm Sciences, 32(1), 1-13.
  8. Barbedo, J. G. A. (2020). Detecting and classifying pests in crops using proximal images and machine learning: A review. AI, 1(2), 312-328.DOI
  9. Battisti, A., and Larsson, S. (2015). Climate change and insect pest distribution range. In Climate change and insect pests (pp. 1-15). CABI Wallingford UK.
  10. Bebber, D. P., Ramotowski, M. A., and Gurr, S. J. (2013). Crop pests and pathogens move polewards in a warming world. Nature Climate Change, 3(11), 985-988.DOI
  11. Benos, L., Tagarakis, A. C., Dolias, G., Berruto, R., Kateris, D., and Bochtis, D. (2021). Machine learning in agriculture: A comprehensive updated review. Sensors, 21(11), 3758.DOI
  12. Bestelmeyer, B. T., Marcillo, G., McCord, S. E., Mirsky, S., Moglen, G., Neven, L. G., Peters, D., Sohoulande, C., and Wakie, T. (2020). Scaling up agricultural research with artificial intelligence. IT Professional, 22(3), 33-38.DOI
  13. Bhatia, A., Chug, A., and Singh, A. P. (2020). Hybrid SVM-LR classifier for powdery mildew disease prediction in tomato plant. In 2020 7th International conference on signal processing and integrated networks (SPIN) (pp. 218-223). IEEE.DOI
  14. Boulent, J., Foucher, S., Théau, J., and St-Charles, P.-L. (2019). Convolutional neural networks for the automatic identification of plant diseases. Frontiers in Plant Science, 10, 941.DOI
  15. Brahimi, M., Boukhalfa, K., and Moussaoui, A. (2017). Deep learning for tomato diseases: classification and symptoms visualization. Applied Artificial Intelligence, 31(4), 299-315.DOI
  16. Brahmi, W., Jdey, I., and Drira, F. (2024). Exploring the role of Convolutional Neural Networks (CNN) in dental radiography segmentation: A comprehensive Systematic Literature Review. Engineering Applications of Artificial Intelligence, 133, 108510.DOI
  17. Brasier, C. (2008). The biosecurity threat to the UK and global environment from international trade in plants. Plant Pathology, 57(5), 792-808.DOI
  18. Chattopadhyay, C., Agrawal, R., Kumar, A., Bhar, L., Meena, P., Meena, R., Khan, S., Chattopadhyay, A., Awasthi, R., and Singh, S. (2005a). Epidemiology and forecasting of Alternaria blight of oilseed brassica in India-a case study/Epidemiologie und Prognose von Alternaria brassicae an Brassica-Ölfrüchten in Indien-Eine Fallstudie. Zeitschrift für Pflanzenkrankheiten und Pflanzenschutz/Journal of Plant Diseases and Protection, 112(4) 351-365. http://www.jstor.org/stable/43215636
  19. Chattopadhyay, C., Agrawal, R., Kumar, A., Singh, Y., Roy, S., Khan, S., Bhar, L., Chakravarthy, N., Srivastava, A., and Patel, B. (2005b). Forecasting of Lipaphis erysimi on oilseed Brassicas in India—a case study. Crop Protection, 24(12), 1042-1053.DOI
  20. Chlingaryan, A., Sukkarieh, S., and Whelan, B. (2018). Machine learning approaches for crop yield prediction and nitrogen status estimation in precision agriculture: A review. Computers and Electronics in Agriculture, 151, 61-69.DOI
  21. Cintra, M. E., Meira, C. A., Monard, M. C., Camargo, H. A., and Rodrigues, L. H. (2011). The use of fuzzy decision trees for coffee rust warning in Brazilian crops. 11th International conference on intelligent systems design and applications, Cordoba, Spain, pp. 1347-1352.DOI
  22. Collier, R. H. (2017). Pest and disease prediction models. In: Thomas, Brian and Murray, Brian G. and Murphy, Denis J., (eds.) Encyclopedia of Applied Plant Sciences: second edition. Waltham, MA: Academic Press, pp. 120-123.DOI
  23. Corrales, D. C., Corrales, J. C., and Figueroa-Casas, A. (2015). Towards Detecting Crop Diseases and Pest by Supervised Learning. Ingeniería y Universidad, 19(1), 207-228.DOI
  24. David, D. (2023). Weather Based Prediction Models for Disease and Pest Using Machine Learning: A Review. Asian Journal of Agricultural Extension, Economics and Sociology, 41(11), 334-345.DOI
  25. de Souza, W. D., Remboski, T. B., de Aguiar, M. S., and Júnior, P. R. F. (2017). A model for pest infestation prediction in crops based on local meteorological monitoring stations. Sixteenth Mexican International Conference on Artificial Intelligence (MICAI), Ensenada, Mexico, pp. 39-45.DOI
  26. Desai, A., Chattopadhyay, C., Agrawal, R., Kumar, A., Meena, R., Meena, P., Sharma, K., Rao, M. S., Prasad, Y., and Ramakrishna, Y. (2004). Brassica juncea powdery mildew epidemiology and weatherbased forecasting models for India—a case study/Die Krankheitsentwicklung des Echten Mehltaus (Erysiphe cruciferarum) auf Brassica juncea und wetterbasierende Modelle zur Vorausschätzung seiner epidemiologischen Entwicklung in Indien—Eine Fallstudie. Zeitschrift für Pflanzenkrankheiten und Pflanzenschutz/Journal of Plant Diseases and Protection, 111(5), 429-438. https://www.jstor.org/stable/43216277
  27. Dhawan, A. K. (2016). Integrated pest management in cotton. In Integrated Pest Management in the Tropics (pp. 499–575). New Delhi, India: New India Publishing Agency.
  28. Domingues, T., Brandão, T., and Ferreira, J. C. (2022). Machine learning for detection and prediction of crop diseases and pests: A comprehensive survey. Agriculture, 12(9), 1350.DOI
  29. Du, J. (2018). Understanding of object detection based on CNN family and YOLO. Journal of Physics: Conference Series, 2nd International Conference on Machine Vision and Information Technology (CMVIT 2018), Hong Kong, 1004, 012029.DOI
  30. Early, R., Bradley, B. A., Dukes, J. S., Lawler, J. J., Olden, J. D., Blumenthal, D. M., Gonzalez, P., Grosholz, E. D., Ibañez, I., and Miller, L. P. (2016). Global threats from invasive alien species in the twenty-first century and national response capacities. Nature Communications, 7(1), 12485.DOI
  31. Emebo, O., Fori, B., Victor, G., and Zannu, T. (2019). Development of tomato septoria leaf spot and tomato mosaic diseases detection device using raspberry Pi and deep convolutional neural networks. Journal of Physics: Conference Series, 1299, 01211810.DOI
  32. Fuentes, A., Yoon, S., Kim, S. C., and Park, D. S. (2017). A robust deep-learning-based detector for real-time tomato plant diseases and pests recognition. Sensors, 17(9), 2022.DOI
  33. Gauriau, O., Galárraga, L., Brun, F., Termier, A., Davadan, L., and Joudelat, F. (2024). Comparing machine-learning models of different levels of complexity for crop protection: A look into the complexity-accuracy tradeoff. Smart Agricultural Technology, 7, 100380.DOI
  34. Girshick, R., Donahue, J., Darrell, T., and Malik, J. (2014). Rich feature hierarchies for accurate object detection and semantic segmentation. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) (pp. 580-587). IEEE.DOI
  35. Gonzalez-Huitron, V., León-Borges, J. A., Rodriguez-Mata, A., Amabilis-Sosa, L. E., Ramírez-Pereda, B., and Rodriguez, H. (2021). Disease detection in tomato leaves via CNN with lightweight architectures implemented in Raspberry Pi 4. Computers and Electronics in Agriculture, 181, 105951.DOI
  36. Gurle, A. S., Barathe, S. N., Gangule, R. S., Jagtap, S. D., and Patankar, T. (2019). Survey paper on tomato crop disease detection and pest management. International Journal of Applied Evolutionary Computation (IJAEC), 10(3), 10-18.DOI
  37. Halstead, M., McCool, C., Denman, S., Perez, T., and Fookes, C. (2018). Fruit quantity and ripeness estimation using a robotic vision system. IEEE robotics and automation LETTERS, 3(4), 2995-3002.DOI
  38. Heeb, L., Jenner, E., and Cock, M. J. (2019). Climate-smart pest management: building resilience of farms and landscapes to changing pest threats. Journal of Pest Science, 92(3), 951-969.DOI
  39. Hernandez-Espinoza, L. H., and Barrios-Masias, F. H. (2020). Physiological and anatomical changes in tomato roots in response to low water stress. Scientia Horticulturae, 265, 109208.DOI
  40. Hu, W., Hong, W., Wang, H., Liu, M., and Liu, S. (2023). A Study on Tomato Disease and Pest Detection Method. Applied Sciences, 13(18), 10063.DOI
  41. Hughes, D., and Salathé, M. (2015). An open access repository of images on plant health to enable the development of mobile disease diagnostics. arXiv preprint arXiv,1511.08060.DOI
  42. Hunt, E. R., Jr. Daughtry, C. S. T., Walthall, C. L., McMurtrey, J. E., and Dulaney, W. P. (2004). Agricultural remote sensing using radio-controlled model aircraft. Digital Imaging and Spectral Techniques: Applications to Precision Agriculture and Crop Physiology. ASA Special Publication no. 66.DOI
  43. Ifft, J., Kuhns, R., and Patrick, K. (2018). Can machine learning improve prediction–an application with farm survey data. International Food and Agribusiness Management Review, 21(8), 1083-1098.DOI
  44. Ilic, M., Ilic, S., Jovic, S., and Panic, S. (2018). Early cherry fruit pathogen disease detection based on data mining prediction. Computers and electronics in agriculture, 150, 418-425.DOI
  45. Isard, S. A., Russo, J. M., Magarey, R. D., Golod, J., and VanKirk, J. R. (2015). Integrated pest information platform for extension and education (iPiPE): progress through sharing. Journal of Integrated Pest Management, 6(1), 15.DOI
  46. Javaid, M., Haleem, A., Khan, I. H., and Suman, R. (2023). Understanding the potential applications of Artificial Intelligence in Agriculture Sector. Advanced Agrochem, 2(1), 15-30.DOI
  47. Jawade, P., Chaugule, D., Patil, D., and Shinde, H. (2020). Disease prediction of mango crop using machine learning and IoT. Advances in Decision Sciences, Image Processing, Security and Computer Vision: International Conference on Emerging Trends in Engineering (ICETE), 1, Springer, Cham.DOI
  48. Jia, S., Gao, H., and Hang, X. (2020). Tomato Pests and Diseases Classification Model Based on Optimized Convolutional Neural Network. Journal of Physics: Conference Series, 1437.DOI
  49. Junjing, L., Leigang, S., and Wenjiang, H. (2019). Research Progress in Monitoring and Forecasting of Crop Diseases and Pests by Remote Sensing. Remote Sensing Technology and Application, 34(1), 21-32.DOI
  50. Kamilaris, A., and Prenafeta-Boldú, F. X. (2018). A review of the use of convolutional neural networks in agriculture. The Journal of Agricultural Science, 156(3), 312-322.DOI
  51. Kattenborn, T., Leitloff, J., Schiefer, F., and Hinz, S. (2021). Review on Convolutional Neural Networks (CNN) in vegetation remote sensing. ISPRS Journal of Photogrammetry and Remote Sensing, 173, 24-49.DOI
  52. Kaur, K., and Kaur, M. (2018). Prediction of plant disease from weather forecasting using data mining. International Journal on Future Revolution in Computer Science and Communication Engineering, 4(4), 685-688. http://www.ijfrcsce.org/index.php/ijfrcsce/article/view/1591
  53. Kenis, M., Rabitsch, W., Auger-Rozenberg, M. A., and Roques, A. (2007). How can alien species inventories and interception data help us prevent insect invasions? Bulletin of Entomological Research, 97(5), 489-502.DOI
  54. Kim, J. A., Sung, J. Y., and Park, S. H. (2020). Comparison of Faster-RCNN, YOLO, and SSD for real-time vehicle type recognition. 2020 IEEE international conference on consumer electronics-Asia (ICCE-Asia).DOI
  55. Lee, Y. S., and Liu, W. Y. (2014). Forecasting value of agricultural imports using a novel two-stage hybrid model. Computers and Electronics in Agriculture, 104, 71-83.DOI
  56. Li, C., Liu, J., Bai, W., Wu, S., Zheng, P., Zhang, J., Pan, Z., and Zhai, J. (2022). Superior energy storage performance in (Bi0.5Na0.5) TiO3-based lead-free relaxor ferroelectrics for dielectric capacitor application via multiscale optimization design. Journal of Materials Chemistry A, 10(17), 9535-9546.DOI
  57. Liebhold, A. M., Brockerhoff, E. G., Garrett, L. J., Parke, J. L., and Britton, K. O. (2012). Live plant imports: the major pathway for forest insect and pathogen invasions of the US. Frontiers in Ecology and the Environment, 10(3), 135-143.DOI
  58. Lima, M. C. F., de Almeida Leandro, M. E. D., Valero, C., Coronel, L. C. P., and Bazzo, C. O. G. (2020). Automatic detection and monitoring of insect pests—A review. Agriculture, 10(5), 161.DOI
  59. Lin, T. L., Chang, H. Y., and Chen, K. H. (2020). The pest and disease identification in the growth of sweet peppers using faster R-CNN and mask R-CNN. Journal of Internet Technology, 21(2), 605-614.DOI
  60. López-Morales, V., López-Ortega, O., Ramos-Fernandez, J., and Munoz, L. (2008). JAPIEST: An integral intelligent system for the diagnosis and control of tomatoes diseases and pests in hydroponic greenhouses. Expert Systems with Applications, 35(4), 1506-1512.DOI
  61. Machekano, H., Mutamiswa, R., Mvumi, B. M., Nyabako, T., Shaw, S., and Nyamukondiwa, C. (2019). Disentangling factors limiting diamondback moth, Plutella xylostella (L.), spatio‐temporal population abundance: A tool for pest forecasting. Journal of Applied Entomology, 143(6), 670-682.DOI
  62. Maduranga, M., and Abeysekera, R. (2020). Machine learning applications in IoT based agriculture and smart farming: A review. International Journal of Engineering Applied Sciences and Technology, 4(12), 24-27.DOI
  63. Mahapatra, S. K., and Dash, A. (2020). ARIMA Model for Forecasting of Black Gram Productivity in Odisha. Asiatic Society for Social Science Research (ASSSR), 2(1), 131-136.DOI
  64. Malicdem, A. R., and Fernandez, P. L. (2015). Rice blast disease forecasting for northern Philippines. WSEAS Transactions on Information Science and Applications, 12, 120-129. Retrieved from https://wseas.com/journals/isa/2015/a225709-430.pdf
  65. Mehta, S., Agrawal, R., and Kumar, A. (2005). Forewarning crop pests and diseases: IASRI methodologies. New Delhi, India: IASRI Publication.
  66. Mehta, S., Agrawal, R., Shukla, R., and Sharma, S. (2001, February). A statistical model for prediction of mango fruit fly outbreak. Paper presented at the National Seminar on Agrometeorological Research for Sustainable Agricultural Production, India.
  67. Mrisho, L. M., Mbilinyi, N. A., Ndalahwa, M., Ramcharan, A. M., Kehs, A. K., McCloskey, P. C., Murithi, H., Hughes, D. P., and Legg, J. P. (2020). Accuracy of a smartphone-based object detection model, PlantVillage Nuru, in identifying the foliar symptoms of the viral diseases of cassava–CMD and CBSD. Frontiers in Plant Science, 11, 590889.DOI
  68. Musolin, D. L. (2007). Insects in a warmer world: ecological, physiological and life‐history responses of true bugs (Heteroptera) to climate change. Global Change Biology, 13(8), 1565-1585.DOI
  69. Nagar, H., and Sharma, R. S. (2020). A comprehensive survey on pest detection techniques using image processing. 4th International Conference on Intelligent Computing and Control Systems (ICICCS), Madurai, India, pp. 43-48.DOI
  70. Ngugi, L. C., Abelwahab, M., and Abo-Zahhad, M. (2021). Recent advances in image processing techniques for automated leaf pest and disease recognition–A review. Information Processing in Agriculture, 8(1), 27-51.DOI
  71. Nihar, F., Khanom, N. N., Hassan, S. S., and Das, A. K. (2021). Plant disease detection through the implementation of diversified and modified neural network algorithms. Journal of Engineering Advancements, 2(01), 48-57.DOI
  72. Pak, M., and Kim, S. (2017). A review of deep learning in image recognition. 4th international conference on computer applications and information processing technology (CAIPT), Kuta Bali, Indonesia, pp. 1-3.DOI
  73. Pandit, P., Krishnamurthy, K., and Bakshi, B. (2022). Prediction of crop yield and pest-disease infestation. In AI, Edge and IoT-based Smart Agriculture, pp. 375-393.DOI
  74. Patel, D., and Bhatt, N. (2021). Improved accuracy of pest detection using augmentation approach with Faster R-CNN. IOP Conference Series: Materials Science and Engineering, 1042, 012020.DOI
  75. Pawara, S., Nawale, D., Patil, K., and Mahajan, R. (2018). Early detection of pomegranate disease using machine learning and internet of things. 3rd International Conference for Convergence in Technology (I2CT). Pune, India, pp. 1-4.DOI
  76. Paymode, A. S., Magar, S. P., and Malode, V. B. (2021). Tomato leaf disease detection and classification using convolution neural network. International Conference on Emerging Smart Computing and Informatics (ESCI), Pune, India, pp. 564-570.DOI
  77. Petrellis, N. (2017). A smart phone image processing application for plant disease diagnosis. 6th international conference on modern circuits and systems technologies (MOCAST), Thessaloniki, Greece, pp. 1-4.DOI
  78. Prajwala, T. M., Pranathi, A., SaiAshritha, K., Chittaragi, N. B., and Koolagudi, S. G. (2018). Tomato leaf disease detection using convolutional neural networks. Eleventh international conference on contemporary computing (IC3), Noida, India, pp. 1-5.DOI
  79. Prashar, N., and Sangal, A. (2022). Plant disease detection using deep learning (convolutional neural networks). Second International Conference on Image Processing and Capsule Networks: ICIPCN, 2021, 2. Lecture Notes in Networks and Systems, vol 300. Springer, Cham.DOI
  80. Priya, R., and Ramesh, D. (2020). ML based sustainable precision agriculture: A future generation perspective. Sustainable Computing: Informatics and Systems, 28, 100439.DOI
  81. Qi, F., Wang, Y., Tang, Z., and Chen, S. (2023). Real-time and effective detection of agricultural pest using an improved YOLOv5 network. Journal of Real-Time Image Processing, 20(2), 33.DOI
  82. Qiang, Z., Shihao, S., Yulin, W., Mengying, L., Hongkai, L., and Qiang, N. (2020). Research on load distribution control technology for parallel operation of power source with different rated capacity. 12th IEEE PES Asia-Pacific Power and Energy Engineering Conference (APPEEC), Nanjing, China, pp. 1-5.DOI
  83. Rajni, J., Sonajharia, M., and Ramasubramanian, V. (2009). Machine learning for forewarning crop diseases. Journal of the Indian Society of Agricultural Statistics, 63(1), 97-107.
  84. Rawat, N., Karnatak, A. K., and Srivastava, R. M. (2020). Population dynamics of okra shoot and fruit borer (Earias vittella) of okra in agro-climatic condition of Pantnagar. International Journal of Chemical Studies, 8(1), 2131-2134.DOI
  85. Rehman, T. U., Mahmud, M. S., Chang, Y. K., Jin, J., and Shin, J. (2019). Current and future applications of statistical machine learning algorithms for agricultural machine vision systems. Computers and Electronics in Agriculture, 156, 585-605.DOI
  86. Ren, S., He, K., Girshick, R., and Sun, J. (2015). Faster R-CNN: Towards real-time object detection with region proposal networks. Advances in Neural Information Processing Systems, 28.DOI
  87. Roques, A., Auger-Rozenberg, M.-A., Blackburn, T. M., Garnas, J., Pyšek, P., Rabitsch, W., Richardson, D. M., Wingfield, M. J., Liebhold, A. M., and Duncan, R. P. (2016). Temporal and interspecific variation in rates of spread for insect species invading Europe during the last 200 years. Biological Invasions, 18, 907-920.DOI
  88. Saeed, H., Ehetisham-ul-Haq, M., Atiq, M., Kamran, M., Idrees, M., Ali, S., Burhan, M., Mohsan, M., Iqbal, M., and Nazir, S. (2018). Prediction of cotton leaf curl virus disease and its management through resistant germplasm and bio-products. Archives of Phytopathology and Plant Protection, 51(3-4), 170-186.DOI
  89. Santini, A., Ghelardini, L., De Pace, C., Desprez‐Loustau, M.-L., Capretti, P., Chandelier, A., Cech, T., Chira, D., Diamandis, S., and Gaitniekis, T. (2013). Biogeographical patterns and determinants of invasion by forest pathogens in Europe. New Phytologist, 197(1), 238-250.DOI
  90. Santini, A., Liebhold, A., Migliorini, D., and Woodward, S. (2018). Tracing the role of human civilization in the globalization of plant pathogens. The ISME journal, 12(3), 647-652.DOI
  91. Sarukhán, J., Whyte, A., Hassan, R., Scholes, R., Ash, N., Carpenter, S., Pingali, P., Bennett, E., Zurek, M., and Chopra, K. (2005). Millennium Ecosystem Assessment: Ecosystems and human well-being. Retrieved from http://www.millenniumassessment.org/en/products.aspx
  92. Setiyowati, S., Nugraha, R. F., and Mukhaiyar, U. (2015). Non-stationary time series modeling on caterpillars pest of palm oil for early warning system. AIP Conference Proceedings, 1, 1692, 020011.DOI
  93. Shorten, C., and Khoshgoftaar, T. M. (2019). A survey on image data augmentation for deep learning. Journal of Big Data, 6(1), 1-48.DOI
  94. Siddiqua, A., Kabir, M. A., Ferdous, T., Ali, I. B., and Weston, L. A. (2022). Evaluating plant disease detection mobile applications: Quality and limitations. Agronomy, 12(8), 1869.DOI
  95. Singh, B., Singh, R., Tiwari, P., and Kumar, N. (2019). Climate based factor analysis and epidemiology prediction for potato late blight using machine learning approaches. Women Institute of Technology Conference on Electrical and Computer Engineering (WITCON ECE), Dehradun, India, pp. 113-122.DOI
  96. Singh, J., Das, D., Vennila, S., and Rawat, K. (2018). Weather based forewarning of pest and disease: An important adaptation strategies under the impact of climate change scenario: A brief review. International Journal of Advanced Multidisciplinary Scientific Research (IJAMSR), 1, 6-21.DOI
  97. Sladojevic, S., Arsenović, M., Anderla, A., Culibrk, D., and Stefanović, D. (2016). Deep neural networks based recognition of plant diseases by leaf image classification. Computational Intelligence and Neuroscience.DOI
  98. Srivastava, R., Gupta, C., Gupta, H., Singh, N., and Kumar, N. (2015). Mathematical modelling of crop yield forecasting and forewarning of pests/diseases. In Proceedings of the International Conference of Advanced Research and Innovation (ICARI) (pp. 417-419.
  99. Sudduth, K. A., Woodward-Greene, M. J., Penning, B. W., Locke, M. A., Rivers, A. R., and Veum, K. S. (2020). AI down on the farm. IT Professional, 22(3), 22-26.DOI
  100. Sykes, J. R., Denby, K. J., and Franks, D. W. (2023). Computer vision for plant pathology: A review with examples from cocoa agriculture. Applications in Plant Sciences, e11559.DOI
  101. Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., and Rabinovich, A. (2015). Going deeper with convolutions. Proceedings of the IEEE conference on computer vision and pattern recognition (CVPR), Boston, MA, USA, pp. 1-9.DOI
  102. Trebicki, P., and Finlay, K. (2019). Pests and diseases under climate change; its threat to food security. In Food Security and Climate Change (pp. 229–249). Wiley.DOI
  103. Veeragandham, S., and Santhi, H. (2020). A review on the role of machine learning in agriculture. Scalable Computing: Practice and Experience, 21(4), 583-589.DOI
  104. Verma, S., and Zhang, Z. L. (2018). Graph capsule convolutional neural networks. arXiv preprint arXiv:1805.08090.DOI
  105. Vishwa Dhar, V. D., Singh, S., Kumar, M., Agrawal, R., and Amrender Kumar, A. K. (2007). Prediction of pod-borer (Helicoverpa armigera) infestation in short-duration pigeonpea (Cajanus cajan) in central Uttar Pradesh. Indian Journal of Agricultural Sciences, 77(10), 701-704.
  106. Wang, X., Yan, Y., and Li, Z. (2022). Machine Learning for Detection and Prediction of Crop Diseases and Pests: A Comprehensive Survey. Agriculture, 12(9), 1350.DOI
  107. Wiesner-Hanks, T., Stewart, E. L., Kaczmar, N., DeChant, C., Wu, H., Nelson, R. J., Lipson, H., and Gore, M. A. (2018). Image set for deep learning: field images of maize annotated with disease symptoms. BMC Research Notes, 11(1), 1-3.DOI
  108. Yang, Z., Xu, C., Wang, M., Zhao, H., Zheng, Y., Huang, H., Vuguziga, F., and Umutoni, M. (2019). Enhancing the thermotolerance of tomato seedlings by heat shock treatment. Photosynthetica, 57(4).DOI
  109. Zhao, Y., Liu, L., Xie, C., Wang, R., Wang, F., Bu, Y., and Zhang, S. (2020). An effective automatic system deployed in agricultural Internet of Things using Multi-Context Fusion Network towards crop disease recognition in the wild. Applied Soft Computing, 89, 106128.DOI
  110. Zhaoxin, G., Han, L., Zhijiang, Z., and Libo, P. (2022). Design a robot system for tomato picking based on YOLO v5. IFAC-PapersOnLine, 55(3), 166-171.DOI
  111. Zheng, Y. Y., Kong, J. L., Jin, X. B., Wang, X. Y., Su, T. L., and Zuo, M. (2019). CropDeep: The Crop Vision Dataset for Deep-Learning-Based Classification and Detection in Precision Agriculture. Sensors, 19(5), 1058.DOI

Authors retain the copyright. This is an open access article distributed under Creative Commons Attribution 4.0 International License (CC BY 4.0)

Send comment about this article
Enter Name.
Enter a valid email address.
Enter a vaid affiliation.
Enter comments (At leaset 10 words)
CAPTCHA Image
Enter Security Code Correctly.

  • Receive Date 14 November 2024
  • Revise Date 15 December 2024
  • Accept Date 08 January 2025
  • First Publish Date 26 August 2025

Home

Editorial Board

Publication Ethics

Contact Us