African Journal of
Agricultural Research

  • Abbreviation: Afr. J. Agric. Res.
  • Language: English
  • ISSN: 1991-637X
  • DOI: 10.5897/AJAR
  • Start Year: 2006
  • Published Articles: 6849

Full Length Research Paper

Artificial intelligence and deep learning based technologies for emerging disease recognition and pest prediction in beans (phaseolus vulgaris l.): A systematic review

Michael Pendo John Mahenge
  • Michael Pendo John Mahenge
  • Department of Informatics and Information Technology, Sokoine University of Agriculture, United Republic of Tanzania.
  • Google Scholar
Hussein Mkwazu
  • Hussein Mkwazu
  • Department of Informatics and Information Technology, Sokoine University of Agriculture, United Republic of Tanzania.
  • Google Scholar
Camilius A. Sanga
  • Camilius A. Sanga
  • Department of Informatics and Information Technology, Sokoine University of Agriculture, United Republic of Tanzania.
  • Google Scholar
Richard Raphael Madege
  • Richard Raphael Madege
  • Department of Crop Science and Horticulture, Sokoine University of Agriculture, United Republic of Tanzania.
  • Google Scholar
Beatrice Mwaipopo
  • Beatrice Mwaipopo
  • Department of Crop Science and Horticulture, Sokoine University of Agriculture, United Republic of Tanzania.
  • Google Scholar
Caroline Maro
  • Caroline Maro
  • Department of Crop Science and Horticulture, Sokoine University of Agriculture, United Republic of Tanzania.
  • Google Scholar


  •  Received: 29 September 2022
  •  Accepted: 15 December 2022
  •  Published: 31 March 2023

 ABSTRACT

Artificial Intelligence (AI) and deep learning have the capacity to reduce losses in crop production, such as low crop yields, food insecurity, and the negative impacts on a country’s economy caused by crop infections. This study aims to find the knowledge and technological gaps associated with the application of AI-based technologies for plant disease detection and pest prediction at an early stage and recommend suitable curative measures. An evidence-based framework known as the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) methodology was used to conduct systematic reviews of the state-of-the-art of AI and deep learning techniques for crop disease identification and pest prediction in developing countries. The results demonstrate that conventional methods for plant disease management face some challenges, such as being costly in terms of labour, having low detection and prediction accuracy, and some are not environmentally friendly. Also, the rapid increase in data-intensive and computational-intensive tasks needed for plant disease classification using traditional machine learning methods poses challenges such as high processing time and storage capacity. Consequently, this paper recommends a deep learning and AI-based strategy to enhance the detection, prediction and prevention of crop diseases. These recommendations will be the starting point for future research.

Key words: Plant diseases detection, pest prediction, pesticide recommendation, artificial intelligence, machine learning.


 INTRODUCTION

The common bean (Phaseolus vulgaris L.) is considered to be the potential grain legume used as food as well as the key source of important micronutrients needed by millions of  people  worldwide,  including  in  sub-Saharan Africa (Said and Taher, 2020). In Tanzania, consumption of common bean dry seeds with cereal-based food guarantees access to a balanced diet and supplementary nutrients  that  help   alleviate   malnutrition   and  prevent various diseases such as cardiovascular, certain types of cancer, and diabetes (Muimui, 2010; Hangen and Benink, 2003). In 2019, about 28.9 million tons of common beans were produced, with tropical low-income countries producing three-quarters of the total (FAOSTAT, 2016). Latin America is the leading bean producer worldwide, with Brazil being a leading country for both bean production and consumer (Rezende et al., 2018). In Africa, Tanzania is the prevalent producer of beans, followed by Uganda and Kenya (Philipo et al., 2020). In Tanzania, the main bean production areas are the northern region, the Great-lakes zone and the southern highlands, where more than 90% are produced by small-scale farmers on farms ranging from 0.5 to 2 ha in size (Ndakidemi et al., 2006). Unfortunately, the importance of the crop cannot be adequately measured due to numerous biotic and abiotic constraints. Biotic constraints refers to the numerous infections caused by bacteria, fungi, viruses, and nematodes (Markell et al., 2012; Pamela et al., 2014; Chilagane et al., 2016; Mpeguzi et al., 2020). Numerous infectious diseases are posing major limitations to bean productivity in Tanzania and worldwide (Hillocks et al., 2006). This in turn threatens the economy of many countries by affecting the production and quality of agricultural products and ultimately reducing food security. Other limiting factors include varieties with low genetic potential for yield; poor soil fertility; weather conditions; drought; insect pests and diseases  (Hillocks   et  al.,  2006).  Some  crop  diseases common in Tanzania include Angular Leaf Spot (ALS) (Phaeoisariopsis griseola), root rots (a complex of pathogens), common bacterial blight (CBB) (Xanthomonas campestris pv. phaseoli), and anthracnose (Colletotrichum lindemuthianum). Other essential diseases include halo blight (Pseudomonas phaseolicola), leaf rust (Uromyces phaseoli), bean common mosaic necrosis virus (BCMNV), and bean common mosaic virus (BCMV) (Mwaipopo et al., 2017). These diseases are prominent in Tanzania, occur in all bean-growing areas and emanates a yield loss of up-to 100% subject to the environment and variety grown (Tryphone et al., 2013). Recently, anthracnose has emerged as a major problem in Tanzania, as observed while conducting a comprehensive survey of common bean viruses in 2016/2017. Moreover, these diseases reduce the quality of harvested-seeds, germination capacity, and market significance, and some can be seed-borne, thereby threatening the seed supply (Degu et al., 2020). Figure 1 displays a synopsis of the symptoms of various diseases in common beans. Bean anthracnose is among the serious infection of beans in Tanzania triggered by the fungal-pathogen called Colletotrichum lindemuthianum (Mpeguzi et al., 2020). If the disease is not well controlled, would cause yield losses of upto 100% (Markell et al., 2012). The infection affects leaves, seeds, pods, and stems of beans. The infected seeds initiate and disperse the disease to the plant (Mudawi et al., 2009).  The  main  symptoms  of  anthracnose  disease in common beans are: growth dark and collapse of veins on the leaf underside; sunken lesions on pods displaying spore creation at the centers; and seed discoloration. Diseased plants may dry and crack, showing seeds inside the pods. Seed formed beneath pod lesions does not show the signs of infection (Padder et al., 2017). In Tanzania, few studies on anthracnose have been indicated regardless of the presence of the disease, such as Padder et al. (2017), Ndee (2013), Shao and Teri (1981). The ALS is the major damaging fungal disease affecting beans (Phaseolus vulgaris L.) spreading in tropical and subtropical areas, causing up-to 80% harvest losses in a given area (Muthomi et al., 2011). The infection is prominent in Eastern and Central African countries, including Tanzania (Pamela et al., 2014; Chilagane et al., 2016). Considering Tanzania as an example, the symptoms has been testified to spread from low to high altitudes above sea level where the common beans are grown (Hillocks et al., 2006) which covers the northern zone, the lake zone, the western regions and the southern regions. ALS has been extensively studied in various previous studies (Wani et al., 2022; Tryphone et al., 2015; Mongi et al., 2016; Mongi, 2018; Chilagane, 2017). Common bean rust is one of the major diseases in the common bean growing areas caused by the fungus called Uromyces appendiculatus (Fromme, 1924; Stavely and Pastor-Corrales, 1989). Leaf rust has been studied in detail by Mmbaga et al. (1996); Liebenberg and Pretorius (2010); Aylor (1990); and Hillock (2006).

The diseases caused by bacteria such as CBB and halo blight are also common worldwide with more cases reported in the USA, Asia, and Africa (Muimui et al., 2011; Chen et al., 2021b; Noble et al., 2019). CBB causes up to 75% of harvest losses (Allen et al., 1996; Muimui et al., 2011). In Tanzania, CBB is very common in all areas that produce beans, especially in the southern highlands of Tanzania. Symptoms of CBB appear on seedlings, foliage, stems, pods, and seeds of common bean plants (Chen et al., 2021b; Wohleb, 2011). CBB and halo blight diseases have been extensively studied to identify symptoms, factors triggering the incidence of disease, and curative measures (Chen et al., 2021a; Tugume et al., 2019; Chen et al., 2021a; Noble et al., 2019; Mwamahonje, 2018; McGrath, 2021). Moreover, common beans are infected with various viral diseases such as BCMV, BCMNV, and Bean Yellow Mosaic Virus (BYMV). In Tanzania, the above-mentioned viruses were discovered in bean samples using the next-generation sequencing method. More studies are needed to address their incidence, distribution, and severity, to complement few reported viruses (Mwaipopo et al., 2018). BCMV and BCMNV are the most common viruses that affect beans globally, where the beans are produced (Worrall et al., 2015). This in return causes up-to 80% harvest losses (Drijfhout, 1991). Consequently, various published articles investigated efficient mechanisms to detect viruses causing bean mosaic disease from the 1980s to date in Tanzania (Mwaipopo  et  al.,  2017;  Mwaipopo  et al., 2018). Plant disease detection and prediction of harmful emerging pests is a challenging task due to the lack or inadequate tools for efficient surveillance, prediction, and prevention. Traditional methods based on scouting in combination with various diagnostic tools are labour and time-intensive. The use of deep learning and emerging artificial intelligence (AI)-based technologies to detect plant diseases and predict harmful pests is an interesting research area that needs to be explored.

This paper contributes to the implementation of the second Sustainable Development Goal (SDG), which focuses on eradicating hunger, guaranteeing food security and sustainable agriculture by 2030. It also addresses the goal of the 3rd SDG, which targets to safeguard healthy lives and stimulate well-being by 2030. Likewise, the aim of the target of the fourth goal is to ensure a well-educated and actively learning society, for which this paper leverages innovative emerging AI-based tools to propose a framework that can detect, classify, and disseminate knowledge on plant disease and pest control measures to ensure food security. This paper reviews new and novel AI-based tools for plant disease detection and prediction of harmful pests at an early stage and to recommend cost-efficient solutions to minimize damage to crop production.


 RESEARCH METHODOLOGY

This study applied the well-known evidence-based framework namely PRISMA to support standard reporting and conduction of systematic reviews (Moher et al., 2009; McKenzie et al., 2021). Four key steps were followed to conduct this study which includes identification of data-sources and search plan, article screening, quality evaluation, and data extraction and analysis. Figure 2 presents the PRISMA workflow for systematic review of the AI-based solutions in crop disease detection and harmful pest prediction.

Identification of data-sources and search strategy

During the period from April 2022 to June 2022, we conducted an extensive literature search in well-known online databases such as Scopus, Science Citation Index (SCI), and other related databases indexing journals such as IEEE, Springer, Elsevier, ACM, and IGI Global Publishers. The keywords considered for searching cut across crop disease management and the application of AI-based technology in agriculture, such as “crop disease classification”, “pest management”, “artificial intelligence”, “Internet-of-Things”, “machine learning” and crop disease management. We considered various sources of information such as articles and reports published in English between 2006 and 2022. Therefore, the search query guiding information retrieval from the databases was (“crop disease classification” OR “crop disease management” OR “pest management” OR “plant disease identification”) AND (“artificial intelligence” OR “Internet-of-Things” OR “machine learning”) AND “precision agriculture”).

Article screening

Initially,  the  total number of identified records based on the guiding search query was 523 records. We conducted a manual evaluation of the retrieved records, considering analysis of various criteria such as title, abstract, index terms, publication language, and duplication. The initial evaluation process identified 69 records to be omitted for exclusion in the screening phase. The remaining 454 records were thoroughly screened based on their context and relevance to our study. The results of the screening process excluded 217 records that were found out of context and irrelevant to our study.

Quality evaluation

The remaining 237 records were considered in the assessment for eligibility to be selected as relevant records. Based on the exclusion criteria, 114 records were found to be irrelevant and were therefore excluded based on the quality issues.

Data extraction and analysis

Finally, 123 records were found relevant and qualified for inclusion in the qualitative synthesis as a sample for this study. A summary of the qualified papers was prepared using Microsoft excel 2013 and refined to filter information such as publication details, techniques used, performance evaluation indicators, dataset used for machine learning, and performance accuracy achieved for each technique. The summarized data were analyzed and reported on the basis of PRISMA checklist (Moher et al., 2009; McKenzie et al., 2021).


 RESULTS AND DISCUSSION

In Tanzania, agriculture is the backbone of the national economy with 80% of Tanzanians considering agriculture as their source of revenue (FAO, 2016). Also, Tanzania is considered to be among the leading countries in producing beans. The area under common bean production has increased from 1961 to 2016, thereby increasing total production. The increased bean production emanates from the growth of the total production fields rather than productivity. Low productivity is due to many diseases, as discussed earlier in this document. The source of diseases and how severely they damage common bean fields are influenced by a number of factors such as host resistance, the vectors, pathogen genetics, and the environment. For example, many crops are threatened by plant-viruses because it is challenging to control viruses because of inadequacy and economically sustainable methods for large production areas (Thresh, 2003). The pot viruses are particularly problematic because they are readily transmitted in seeds, so they may move over long distances in this form without being spread in the immediate environment by insect vectors. This leads to difficulties in their management (Galves and Molares, 1989; Shukla and Ward, 1989). Some fungal and bacterial diseases are transmitted in and on seeds, and all may be harboured in plant debris and spread by rainfall splashes. These varied modes of transmission make it difficult to control these diseases. Often, an understanding of the seasonal cycles of hostplants, pathogens, and vectors is needed in order to devise proper control measures for a disease. Different strategies can be used to minimize infections of common beans by plant pathogens depending on the nature of the disease, as described in the following section.

Conventional plant disease management methods

Conventional plant disease management methods include various strategies such as Integrated Pest Management (IPM), cultural, physical, chemical, and biological controls, and host plant resistance. IPM is an active and environmentally sensitive approach to pest management that involves a combination of common- sense practices. IPM plans use up-to-date, comprehensive evidence on pest life cycles and their interactions or relationships with the environment. This information is used in combination with available pest control methods to manage pest damage by the most economical means and with the least possible hazard to people, property, and the environment. Disease control is the most effective method through the use of IPM; crop rotation; controlling of weeds; destroying of old crops; avoiding planting of new crops on diseased plantings; and rouging are very important (Persley et al., 2010; Munir, 2017). However, the IPM method is labour-intensive, as reported by Jørs et al. (2017). The cultural method is the method of controlling bean diseases by removing or burying infected plant debris after harvest to reduce overwinter survival of the pathogen (Mohammed, 2013b). It also includes two to three years of crop rotation with non-host plants to break the disease cycle and minimise the chances of pathogen survival (Coyne et al., 2003). Weekly scouting of the field for disease symptoms can assist in such a way that infected plants can be uprooted and removed. Additional cultural controls may include adequate plant spacing and removal of weeds to allow air circulation and foliar drying, as cool, humid conditions can promote certain diseases (McGee, 1995). Overhead irrigation should be avoided because it can cause the leaves to be wet, resulting in pathogen proliferation and sporulation on the foliage. Alternatively, drip irrigation that supplies water to the plant at the root zone (Buruchara et al., 2010) does not wet the leaves and can reduce disease incidence. Moreover, sanitation is another cultural method that has been implemented through the use of certified disease-free seeds and use of IPM. Physical methods include heat treatment through soil solarization with a plastic sheet over the plot about a month before planting, which reduces the incidence of pathogens such as anthracnose on beans. Apart from physical methods, there are biological methods that involve some fungi and bacteria producing bioactive volatile organic compounds that inhibit the growth of pathogens.  For example, Mota et al. (2021) reported using endophytic fungi belonging to the genus Induratia spp. to control anthracnose, angular leaf spot, and white mould in beans. The effectiveness of bioagents against bean-causing disease pathogens was also supported by Abdel-Fattah et al. (2011), who revealed that the use of arbuscular mycorrhizal fungi as bioagents in common bean reduced the incidence and severity of Rhizoctonia root rot. Therefore, from these studies, it shows that different bioagents can be used as biological controls against bean disease-causing pathogens and they can offer a promising control of the diseases, although physical and biological methods still have weaknesses by Btryon (2022). Chemical methods comprise specific pesticides and have been proven to offer control of bean disease pathogens. Pesticides that are effective against Pythium spp have been reported by Abawi et al. (2006).

Also, the list of chemicals that are effective in reducing the severity and incidence of bean anthracnose, resulting in increased yield, has been reported (Mohammed et al., 2013a; Beshir, 2003). Moreover, phosphoric acid, benzoic acid, Bion, and pyrocatechol applications offered good control of Uromyces appendiculatus, a pathogen causing bean rust (Mansour et al., 2016). Lemessa et al. (2011) recommended spraying Benomyl for effective management of angular leaf spot.

Furthermore, various chemicals such as copper sulphate, copper hydroxide, and potassium methy-dithiocarbamate have been reported to control foliage infection of bean common bacterial blight (Karavina et al., 2011). Apart from other methods, host plant resistance is the most efficient and potentially durable disease management strategy for both resource-poor farmers and medium-sized small-holder farmers in Africa. Breeding for resistance is the most affordable method for these resource-poor farmers as they use it to retain their seeds for subsequent cropping cycles. Different cultivars with resistance to different diseases have been screened and developed by breeders as a long-term control of plant diseases. This is the most effective method of controlling viral diseases (Kelly et al., 1995). For example, it has been reported that the dominant I gene and the bc-3 recessive gene together in the same variety of common bean give complete resistance to BCMV and BCMNV (Drijfhout, 1978; Vallejos et al., 2006; Naderpour et al., 2010). Anthracnose and angular leaf spot resistance were explained by Gonçalves-Vidigal et al. (2020). CBB resistance (Miklas et al., 2000), ALS (Oblessuc et al., 2012; Caixeta et al., 2005). Wasonga et al. (2010) developed snap bean lines with broad spectrum rust resistance and heat tolerance for tropical agroecosystems. Offering resistance to beans is crucial in minimising the diseases, though the challenge is to breed the varieties with resistance that is effective, stable, and broad–spectrum (Nelson et al., 2018). Palloix et al. (2009) also reported that resistance breakdown is obvious, and sometimes there is a negative correlation between yield and disease resistance variables, whereby the durability of plant major resistance genes to pathogens depends on the genetic background, experimental evidence, and consequences for breeding strategies. For instance, the wheat rust resistance gene reduces grain yield by 5%. Apart from that, since the resistance involves multiple gene introgression, the planning needs greater effort than single gene resistance. Lastly, horizontal resistance is durable but difficult to relate to an accurate and reliable assessment of the level of resistance.  

Research trends on plant disease management using artificial intelligence and deep learning based technologies

Sensing technology, Unmanned Aerial Vehicle (UAV), Internet of Things (IoT), machine learning techniques and Mobile Cloud Computing (MCC) are driving smart farming and innovative disease management strategies. The following subsection discusses the cost-efficient innovative technologies for efficient management of plant diseases.

Low-cost sensors

Rapid advances in sensor technology have aided in the detection of plant diseases and pests, overcoming the limitations of traditional methods and the reliance on human experts. For example, Huang et al. (2014) analysed the detection of Maize Chlorotic Mottle Virus (MCMV) using a bio-sensor implemented on the basis of Quartz Crystal Microbalance (QCM), which achieved high detection accuracy. Jócsák et al. (2019) investigated the application of Electrochemical Impedance Spectroscopy (EIS) sensors to detect viruses or pathogens that threaten plants. The EIS proved useful, especially for in-field investigations due to its portability and capability to deliver quick feedback. Khater et al. (2019) reported an innovative sensor to detect Citrus Tristeza Virus (CTV) in multiple infections, which is a common scenario for planted crops. Also, wearable sensors play a great role in real-time monitoring of plants’ health. For example, Nassar et al. (2018) described wearable sensors that were configured to monitor plants’ health. The sensors gathered data such as temperature, strain, and humidity that were used to investigate the influence of environmental conditions on the health of plants. Wearable sensors are also widely used in precision agriculture to observe variations in solute content in plants and keep records of water usage in plants (Coppedè et al., 2017).

Unmanned aerial vehicle

The advancement in sensing technology such as Hyperspectral Imaging Sensors (HIS) is augmenting the traditional tools used for capturing airborne and satellite images. Analysing airborne or satellite-based data is expensive and time-consuming (Behmann et al., 2018). Hyperspectral sensing technology such as Unmanned Aerial Vehicles (UAVs) can guarantee a cost-efficient and effective way of acquiring imagery data from the field. UAVs can collect imagery data at a small aerospace attitude and provide an efficient platform for hyperspectral imaging, which is more suitable and cost-effective than traditional methods (Ishengoma et al., 2022). Moreover, UAVs contribute largely to supporting precision farming. For instance, UAVs are widely used in soil and field analysis to improve crop yield prediction through the implementation of data-driven techniques while improving fertilizer and pesticide utilisation (Peña et al., 2013; Torres-Sanchez et al., 2013). Also, crop and spot spraying is another application area of the UAVs that has been reported to be efficient in terms of cost and time (Yallappa et al., 2017; Hentschke et al., 2018; Xiongkui et al., 2017).  Furthermore, Guo et al. (2012) demonstrated the applicability of UAVs in crop monitoring to assist farmers in tracking crop status and identifying areas that require immediate intervention to improve crop conditions.  This study uses the advanced capabilities of UAV-based data to propose an innovative and smart platform for crop disease detection and prediction of harmful pests.

Internet of things technologies

Advances in the Internet of Things (IoT) are transforming many sectors, including agriculture, to provide modern technologies in the agricultural production value chain such as precision farming, automated production processes, yield prediction, remote monitoring of plant diseases and pests, etc. In supporting smart farming, modern and innovative technologies such as IoT, remote sensing, wireless communication, and cloud computing have the potential to achieve real-time monitoring of plant health and the conditions necessary for plant growth (Maksimovic et al., 2017; Bastiaanssen et al., 2000; Hashem et al., 2015; Weber and Romana, 2010). Furthermore, the IoT and UAVs are widely used to track the occurrence of plant diseases and harmful pests. IoT equipment can collect real-time data such as weather data and plant growth by using cost-efficient sensor nodes, while UAVs could take field/farm and crop images that can be analysed to detect the prevalence of crop diseases and harmful pests. For instance, Gao et al. (2020) proposed a framework established on the basis of IoT and UAV to monitor pests and crop diseases. The proposed framework, in particular, considered the relationship between weather and the occurrence of pests and crop diseases. 

Machine learning techniques

Over the years, Machine Learning (ML) techniques have been considered important for crop disease detection. Traditional ML methods are widely used for disease detection; for instance, Support Vector Machine (SVM) and random forest in tomato (Govardhan and Veena, 2019; Mokhtar et al., 2015) and K-Nearest-Neighbors (KNN) in soybean (Shrivastava et al., 2017). An image segmentation algorithm established based on the Genetic Algorithm (GA) has been proposed to detect plant leaf diseases and classify them into appropriate classes (Singh and Misra, 2017). The proposed method used the K-means clustering technique with the lowest distance condition, followed by SVM classification. This, in turn, improved classification accuracy by 9.17%. Nevertheless, traditional ML approaches still face some challenges that need to be addressed, such as being highly dependent on feature extraction needed to train the models, which consumes a lot of time, and ML models are suitable for some specific environments, especially when used to process images (Nigam and Jain, 2020; Shrivastava and Pradhan, 2021). To address the aforementioned challenges, deep learning algorithms with automated feature extraction are promising solutions for crop disease detection and classification (Agarwal et al., 2020; Kamilaris and Prenafeta, 2018). Table 1 summarizes research trends on deep learning techniques for crop disease detection.

Mobile cloud computing

Mobile Cloud Computing (MCC) technology has the potential to implement an integrated and collaborative solution for automated crop disease detection, tracking, and prediction. Farmers can instantly and precisely detect the prevalence of diseases and acquire recommendations for curative actions with the help of a cloud-based-mobile application such as a real-world mobile-cloud based plant disease diagnosis system. However, the emerging trends of big data have brought substantial challenges in terms of processing time and cost (Hashem et al., 2015; Skourletopoulos et al., 2017; Pallis, 2017). This is due to the advancement of technologies for collecting data from fields such as sensors, UAVs, and mobile devices. To efficiently process this data, in line with meeting Quality of Experience (QoE) the combined and collaborative capabilities of Mobile Edge Computing and Cloud Computing technologies need to be explored. According to the findings of this study, artificial intelligence (AI) and deep learning algorithms are suitable for processing large amounts of data using cloud-based image processing techniques. 

Research gap

Analysis of reviewed papers demonstrates that conventional plant disease management strategies such as IPM are considered to be a tool kit approach to crop protection.  However, the methods prove to be labour-intensive and not easily disseminated to farmers (Jørs et al., 2017). Btryon (2022) and Nzungize et al. (2012) also reported that biological methods of controlling pests or diseases can sometimes fail in their specificity when the predator used a bio-control agent may switch to a different target. It was further reported that the biological methods are slow and do not destroy the pest, only reduce the pathogen. Generally, the application of chemicals in controlling pathogens causing bean diseases has been an efficient method.  However there are environmental concerns about their use and some are no longer available. Chemical treatments for large-scale farmers could contaminate soil and water, whereas small-scale, poor-resource farmers cannot afford to apply chemical control and face health risks associated with handling chemical pesticides due to a lack of education in handling chemical pesticides to the farmer community. Consequently, smart crop disease detection and pest prediction technologies are considered as important worldwide. However, developing countries face numerous challenges in implementing automated systems for crop disease identification and pest prediction. Such challenges include inefficient internet connectivity, high costs for purchasing equipment and deploying smart technology on farms, and a shortage of experts on smart farming technology. Several previous studies proposed ML-based techniques to address the challenges related to plant disease identification, classification, and prediction. However, traditional ML techniques face challenges due to overwhelming progress in resource-intensive and data-driven tasks such as large datasets of images originating in the field. For instance, traditional ML relies heavily on feature extraction for model training, which consumes significant processing time and memory when processing large image datasets (Nigam and Jain, 2020; Shrivastava and Pradhan, 2021). To address this technological gap, this study reviewed various deep learning and AI-based models with automated feature extraction suitable for implementing cost-effective crop disease detection, classification, and prediction solutions. According to the results of the systematic review of this study, the combined strategies of machine learning, recommendation systems, WebGIS and image processing techniques have not been sufficiently explored to implement an AI-based solution for joint disease detection and prediction

Theoretical framework

Figure 3 shows the main building block of the deep learning and AI-based framework for detecting crop diseases and predicting harmful pests. The white-labelled rectangles indicate the existing machine learning strategies, and the directed arrows indicate information workflow from one machine learning task to another. The coloured rectangles indicate innovative tasks proposed by this study for crop disease identification and harmful pest prediction strategies. For instance, the orange-rectangles represent the proposed improvement of the machine learning tasks to predict harmful pests related to the detected disease, while the green-rectangle represents the proposed task that will be implemented in the system recommendation module to recommend possible solutions for the farmer to take curative actions. The proposed framework is expected to be applied by private and government agencies at district, regional, and national levels to monitor the plant’s health at various stages of growth. The trained crop disease detection and harmful pest prediction models map the locations where crops are planted and deploy the AI-based models to detect diseases, predict harmful pests, and recommend possible curative measures. Moreover, this study helps to fill the policy research gap by combining existing and emerging technologies for predicting harmful pests and detecting plant diseases using artificial intelligence technologies in collaboration with crop disease management centers at the community, district, and regional level. Also, the demonstration of the real-world applications at department level, university management level, government officials, private sectors, media, and social media will serve as evidence to engage policymakers. In collaboration with other stakeholders, it is crucial to develop and implement an interactive and social mobilisation strategy to influence decision makers and other stakeholders to mobilise resources needed to deploy the technology in the real work environment. Therefore, a collaborative strategy guarantees public awareness and consent. This, in turn, will ensure the social commitment among all stakeholders in fighting against prevalence of plant diseases and harmful pests in Tanzania.


 CONCLUSION AND FUTURE WORKS

This paper presents a systematic review of the state-of-the-art of artificial intelligence and deep learning techniques for crop disease detection and harmful pest prediction. Initially, the paper presented an overview of the current situation and symptoms of various diseases, specifically fungal diseases, bacterial diseases, and viral diseases affecting crops, particularly common beans. Then, we reviewed various articles on conventional plant disease management strategies to identify areas of improvement. The analysis of the reviewed articles reveals that conventional plant disease management strategies such as IPM are commonly used to minimise the prevalence of plant diseases. However, the review results further identified some challenges related to conventional methods, such as being costly in terms of labour, having low identification and prediction accuracy, and some not being environmentally friendly. Other identified challenges for implementing automated systems for crop disease identification and pest prediction, particularly in developing countries, include inefficient internet connectivity, high costs of acquiring equipment and deploying smart technology in farms, and a shortage of experts in smart farming technology. Previous studies also revealed that machine learning-based techniques were commonly used to process and handle real-time data captured to detect crop diseases. However, in recent years, the rapid increase in data-intensive and computational-intensive demands has posed challenges to traditional machine learning-based methods in terms of server capacity and performance, which should be addressed through complementary technologies capable of processing massive amounts of data, such as cloud computing, mobile cloud computing, edge computing, and big data processing architectures. Therefore, this systematic review recommends a deep learning and AI-based technique with automated feature extraction suitable for implementing cost-efficient crop disease detection, classification, and prediction. Moreover, this paper presents the theoretical framework describing the main components of deep learning and AI-based strategies for crop disease identification and harmful pest prediction. Future work may consider extending and implementing deep learning and AI-based models for crop disease identification and pest prediction in real-working environments. Furthermore, another significant future research direction could be to conduct rigorous usability assessment and design requirements for implementing the solution in the real-working environment of Mobile Edge Computing (MEC) to enable service delivery proximate to users.


 CONFLICT OF INTERESTS

The author has not declared any conflict of interests.


 ACKNOWLEDGEMENT

The work was supported by the Sokoine University of Agriculture (SUA) under SUA Research Innovation Support (SUARIS2) grant (DPRTC/R/126/CoNAS/5/ 2022). Any opinions, findings, and conclusions are those of the authors and do not necessarily reflect the views of the above agency.



 REFERENCES

Abawi GS, Ludwig JW, Gugino BK (2006). Bean root rot evaluation protocols currently used in New York. Annual Report- Bean Improvement Cooperative 49:83-84. 

 

Abdel-Fattah GM, El-Haddad SA, Hafez EE, Rashad YM (2011). Induction of defense responses in common bean plants by arbuscular mycorrhizal fungi. Microbiological Research 166(4):268-281.
Crossref

 

Agarwal M, Gupta SK, Biswas KK (2020). Development of Efficient CNN model for Tomato crop disease identification. Sustainable Computing: Informatics and Systems 28:100407. 
Crossref

 

Ahmad I, Hamid M, Yousaf S, Shah ST, Ahmad MO (2020). Optimizing pretrained convolutional neural networks for tomato leaf disease detection. Complexity 6 p. 
Crossref

 

Allen DJ, Ampofo JKO, Wortmann CS (1996). Pests, disease and nutritional disorders of the common bean in Africa: A field guide. Cali, Colombia: International Centre for Tropical Agriculture; Wageningen, The Netherlands: Technical Centre for Agricultural and Rural Co-operation P 37.

 

Aylor DE (1990). The role of intermittent wind in the dispersal of fungal pathogens. Annual Review of Phytopathology 28(1):73-92. 
Crossref

 

Beshir T (2003). Biology and control of bean anthracnose in Ethiopia. A PhD. Thesis submitted to the Faculty of Natural and Agricultural Sciences, University of Free State. Bloemfontein, South Africa P 144.

 

Btryon (2022). Advantages and Disadvantages of Biological Control.

View

 

Buruchara R, Mukankusi C, Ampofo K (2010). Bean Diseases and Pest Identification and Management. Kampala, UG: International Center for Tropical Agriculture (CIAT); Pan-African Bean Research Alliance (PABRA)-Handbook for Small Scale Seed Producers, 79p.

 

Caixeta ET, Borém A, Alzate-Marin AL, Fagundes SDA, Silva MGDM, de Barros EG, Moreira MA (2005). Allelic relationships for genes that confer resistance to angular leaf spot in common bean. Euphytica 145(3):237-245. 
Crossref

 

Chen J, Chen J, Zhang D, Nanehkaran YA, Sun Y (2021). A cognitive vision method for the detection of plant disease images. Machine Vision and Applications 32(1):1-18. 
Crossref

 

Chen NW, Ruh M, Darrasse A, Foucher J, Briand M, Costa J, Studholme DJ, Jacques MA (2021). Common bacterial blight of bean: a model of seed transmission and pathological convergence. Molecular Plant Pathology 22(12):1464-1480. 
Crossref

 

Chilagane LA (2017). Genotype-pathogen characterization and markers identification for angular leaf spot disease resistance in common bean in Tanzania (Doctoral dissertation, Sokoine University of Agriculture) 82 p.

View

 

Chilagane LA, Nchimbi-Msolla S, Kusolwa PM, Porch TG, Diaz LMS, Tryphone GM (2016). Characterization of the common bean host and Pseudocercospora griseola, the causative agent of angular leaf spot disease in Tanzania. African Journal of Plant Science 10(11):238-245. 
Crossref

 

Coyne DP, Steadman JR, Godoy-Lutz G, Gilbertson R, Arnaud-Santana E, Beaver JS, Myers JR (2003). Contributions of the Bean/Cowpea CRSP to management of bean diseases. Field Crops Research 82(2-3):155-168. 
Crossref

 

Degu T, Yaregal W, Gudisa T (2020). Status of common bean (Phaseolus vulgaris L.) diseases in Metekel Zone, northwest Ethiopia. Journal of Plant Pathology and Microbiology 11(5):494. 
Crossref

 

Drijfhout E (1978). Genetic interaction between Phaseolus vulgaris and bean common mosaic virus with implications for strain identification and breeding for resistance. Agricultural Research Report 872. Wageningen, The Netherlands 98 p.

 

Drijfhout E (1991). Bean common mosaic virus. In: Compendium of bean diseases. Hall R. (Ed.). APS Press, The American Phytopathological Society, Minnesota pp. 37-39.

 

Elaraby A, Hamdy W, Alruwaili M (2022). Optimization of deep learning model for plant disease detection using particle swarm optimizer. Computers, Materials and Continua 71(2):4019-4031.
Crossref

 

FAO (2016) Food and Agriculture Organization of the United Nations statistics division

View

 

Fromme FD (1924). The rust of cowpeas. Phytopathology 14:67-79.

 

Gadekallu TR, Rajput DS, Reddy M, Lakshmanna K, Bhattacharya S, Singh S, Jolfaei, A, Alazab M (2021). A novel PCA-whale optimization-based deep neural network model for classification of tomato plant diseases using GPU. Journal of Real-Time Image Processing 18(4):1383-1396. 
Crossref

 

Galves GE, Morales FJ (1989). Aphid-transmitted viruses. In: Schwartz HF Marcial A, Pastor-Corrales (Eds.) Bean production problems in the Tropics. Centro Internacional de Agricultura Tropical Cali, Colombia pp. 333-342.

 

Gao D, Sun Q, Hu B, Zhang S (2020). A framework for agricultural pest and disease monitoring based on Internet-of-Things and Unmanned Aerial Vehicles. Sensors 20(5):1487. 
Crossref

 

Gonçalves-Vidigal MC, Gilio TAS, Valentini G, Vaz-Bisneta M, Vidigal Filho PS, Song Q, Oblessuc PR, Melotto M (2020). New Andean source of resistance to anthracnose and angular leaf spot: Fine-mapping of disease-resistance genes in California Dark Red Kidney common bean cultivar. PloS one 15(6):e0235215. 
Crossref

 

Govardhan M, Veena MB (2019). Diagnosis of tomato plant diseases using random forest. Global Conference for Advancement in Technology (GCAT) pp. 1-5. 
Crossref

 

Guo T, Kujirai T, Watanabe T (2012). Mapping crop status from an unmanned aerial vehicle for precision agriculture applications. ISPRS-International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences 39:485-490. 
Crossref

 

Hangen L, Bennink MR (2003). Consumption of black beans and navy beans (Phaseolus vulgaris L.) reduced azoxymethane-induced colon cancer in rats. Nutrition and Cancer 44(1):60-65. 
Crossref

 

Hashem IAT, Yaqoob I, Anuar NB, Mokhtar S, Gani A, Khan SU (2015). The rise of big data on cloud computing: Review and open research issues. Information Systems 47:98-115. 
Crossref

 

Hentschke M, Freitas E, Hennig C, Veiga C (2018). Evaluation of altitude sensors for a crop spraying drone. Drones 2(3):25. 
Crossref

 

Hillocks RJ, Madata CS, Chirwa R, Minja EM, Msolla S (2006). Phaseolus bean improvement in Tanzania, 1959-2005. Euphytica 150(1):215-231. 
Crossref

 

Hughes DP, Salathé M (2015). An open access repository of images on plant health to enable the development of mobile disease diagnostics through machine learning and crowdsourcing. CoRR. 2015;abs/1511.08060.

View

 

Ishengoma FS, Rai IA, Ngoga SR (2022). Hybrid convolution neural network model for a quicker detection of infested maize plants with fall armyworms using UAV-based images. Ecological Informatics 67:101502. 
Crossref

 

Jørs E, Aramayo A, Huici O, Konradsen F, Gulis G (2017). Obstacles and opportunities for diffusion of integrated pest management strategies reported by Bolivian small-scale farmers and agronomists. Environmental Health Insights 
Crossref

 

Jócsák I, Végvári G, Vozáry E (2019). Electrical impedance measurement on plants: a review with some insights to other fields. Theoretical and Experimental Plant Physiology 31:359-375.
Crossref

 

Kamilaris FX, Prenafeta-Bold (2018). Deep learning in agriculture: A survey. Computers and Electronics in Agriculture 147:70-90.
Crossref

 

Karavina C, Mandumbu R, Parwada C, Tibugari H (2011). A review of the occurrence, biology and management of common bacterial blight. Journal of Agricultural Technology 7(6):1459-1474.

 

Karthik R, Hariharan M, Anand S, Mathikshara P, Johnson A, Menaka R (2020). Attention embedded residual CNN for disease detection in tomato leaves. Applied Soft Computing Journal 86:105933. 
Crossref

 

Kelly JD, Afanador L, Haley SD (1995). Pyramiding genes for resistance to bean common mosaic virus. Euphytica 82(3):207-212.
Crossref

 

Krishnan VG, Deepa JR, Rao PV, Divya V, Kaviarasan S (2022). An automated segmentation and classification model for banana leaf disease detection. Journal of Applied Biology and Biotechnology 10(1):213-220. 
Crossref

 

Lemessa F, Sori W, Wakjira M (2011). Association between angular leaf spot (PhaeoIsariopsis griseola (Sacc.) Ferraris and common bean (Phaseolus vulgaris) yield at Jimma, southwestern Ethiopia. Plant Pathology Journal 10 (2):57-65.
Crossref

 

Liebenberg MM, Pretorius ZA (2010). Common Bean Rust: Pathology and Control. Published by John wiley and Sons, Inc., Hoboken, New Jersey. Horticultural Reviews 37:1-59.
Crossref

 

Mansour M, El-Fawy, Kamal AM, Abo-Elyousr (2016). Efficacy of certain chemical compounds on common bean rust disease. Archives of Phytopathology and Plant Protection 49 (19-20): 522-532. 
Crossref

 

Markell S, Wunsch M, Del Río L (2012). Anthracnose of Dry Beans. North Dakota State University Extension Publication, Fargo 33p.

 

McGee DC (1995). Epidemiological approach to disease management through seed technology. Annual Review of Phytopathology 33(1):445-466. 
Crossref

 

McGrath MT (2021) Bacterial Diseases of Beans.

View

 

McKenzie M, Bossuyt P, Boutron I, Hoffmann T, Mulrow C (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. Systematic reviews 10(1):1-11 
Crossref

 

Meen TH (2019). IoT, communication, and engineering. In: 2019 IEEE Eurasia Conference on IoT, Communication, and Engineering (IEEE ECICE 2019), vol. 3-6, Yunlin, Taiwan, October 2019.

 

Miklas PN, Smith JR, Riley R, Grafton KF, Singh SP, Jung G, Coyne DP (2000). Marker-assisted breeding for pyramided resistance to common bacterial blight. Bean Improvement Cooperative Annual Report 43:39-40.

View

 

Mmbaga MT, Steadman JR., Eskridge KM (1996). Virulence patterns of Uromyces appendiculatus from different geographical areas and implications for finding durable resistance to rust of common bean. Phytopathology 144(11-12):533-541. 
Crossref

 

Mohammed A (2013a). An overview of distribution, biology and the management of common bean anthracnose. Journal of Plant Patholology and Microbiology 4(8):1-6
Crossref

 

Mohammed A, Ayalew A, Dechassa N (2013b). Effect of integrated management of bean anthracnose (Colletotrichum lindemuthianum Sacc. and Magn.) through soil solarization and fungicide applications on epidemics of the disease and seed health in Hararghe Highlands, Ethiopia. Journal of Plant Pathology and Microbiology 4:182. 
Crossref

 

Moher D, Liberati A, Tetzla, J, Altman DG (2009). Preferred Reporting Items for Systematic Reviews and Meta-Analyses: The PRISMA Statement. Annals of Internal Medicine 151(4):264-269. 
Crossref

 

Mokhtar U, Ali MA, Hassenian AE, Hefny H (2015). Tomato leaves diseases detection approach based on Support Vector Machines. 11th International Computer Engineering Conference (ICENCO), pp. 246-250. 
Crossref

 

Mongi R, Tongoona P, Shimelis H, Sibiya J (2016). Appraisal of common bean farming systems under angular leaf spot disease prone environments of the Southern Highlands of Tanzania. Indian Journal of Agricultural Research 50(5):428-433. 
Crossref

 

Mongi R, Tongoona P, Shimelis H, Sibiya J (2018). Agronomic performance and economics of yield loss associated with angular leaf spot disease of common bean in the southern highlands of Tanzania. Plant Disease 102(1):85-90. 
Crossref

 

Mota SF, Pádua PF, Ferreira AN, Gomes LDBW, Dias MA, Souza EA, Pereira OL, Cardoso PG (2021). Biological control of common bean diseases using endophytic Induratia spp. Biological Control 159:104629. 
Crossref

 

Mpeguzi M, Susan NM, Mabagala R, Chilagane LA (2020). Races of Colletotrichum lindemuthianum (Sacc. Magnus) Briosi Cavara in major bean growing regions in Tanzania. African Journal of Plant Science 14(8):308-314. 
Crossref

 

Mudawi HI, Idris MO, El-Balla MA (2009). Anthracnose disease in common bean (Phaseolus vulgaris L.) in Shambat, Sudan: Disease incidence, severity and effect on yield. University of Khartoum Journal of Agricultural Sciences (Sudan) 16(3):477-489.

 

Muimui KK (2010). Beans stakeholder consultative workshop. commonwealth youth programme Africa. Lusaka-Zambia, January 21-22.

 

Muimui KK, Kimani PM, Muthom JW (2011). Resistance and inheritance of common bacterial blight in yellow bean. African Crop Science Journal 19(4):277-287.

 

Munir M (2017). Management of plant virus diseases; farmer's knowledge and our suggestions. Hosts and Viruses 4:28-33. 

 

Muthomi JW, Muimui KK, Kimani PM (2011). Inheritance of resistance to angular leaf spot in yellow beans. African Crop Science Journal 19(4):267-275.

 

Mwaipopo B, Nchimbi-Msolla S, Njau P, Tairo F, William M, Binagwa P, Kweka E, Kilango M, Mbanzibwa D (2017). Viruses infecting common bean (Phaseolus vulgaris L.) in Tanzania: A review on molecular characterization, detection and disease management options. African Journal of Agricultural Research 12(18):1486-1500. 
Crossref

 

Mwaipopo B, Nchimbi-Msolla S, Njau PJ, Mark D, Mbanzibwa DR (2018). Comprehensive surveys of bean common mosaic virus and bean common mosaic necrosis virus and molecular evidence for occurrence of other Phaseolus vulgaris viruses in Tanzania. Plant Disease 102(11):2361-2370. 
Crossref

 

Mwamahonje A (2018). Screening of local and improved bean varieties for resistance to halo blight disease. African Journal of Agricultural Research 13(32):1667-1673. 
Crossref

 

Ndakidemi PA, Dakora FD, Nkonya EM, Ringo D, Mansoor H (2006). Yield and economic benefits of common bean (Phaseolus vulgaris) and soybean (Glycine max) inoculation in northern Tanzania. Australian Journal of Experimental Agriculture 46(4):571-577. 
Crossref

 

Ndee AK (2013). Introgression of bean anthracnose resistance genes in common bean (Phaseolus vulgaris l.) lines with ALS, CBB and BCMV/BCMNV diseases resistances (Doctoral dissertation, Sokoine University of Agriculture). 59p.

View

 

Nelson R, Wiesner-Hanks T, Wisser R, Balint-Kurti P (2018). Navigating complexity to breed disease-resistant crops. Nature Reviews Genetics 19(1):21-33. 
Crossref

 

Nigam S, Jain R (2020). Plant disease identification using deep learning: A review: Indian Journal of Agricultural Sciences 90(2):249-257.
Crossref

 

Nithish EK, Kaushik M, Prakash P, Ajay R, Veni S (2020). Tomato leaf disease detection using convolutional neural networks. In2018 Eleventh International Conference on Contemporary Computing pp. 1125-1132, Coimbatore, India. 
Crossref

 

Noble TJ, Young AJ, Douglas CA, Williams B, Mundree S (2019). Diagnosis and management of halo blight in Australian mungbeans: A review. Crop and Pasture Science 70(3):195-203. 
Crossref

 

Nzungize JR, Lyumugabe F, Busogoro, JP, Baudoin JP (2012). Pythium root rot of common bean: biology and control methods. Biotechnology, Agronomy, Society and Environment 16(3):405-413.

 

Oblessuc PR, Baroni RM, Garcia AAF, Chioratto AF, Carbonell SAM, Camargo LEA, Benchimol LL (2012). Mapping of angular leaf spot resistance QTL in common bean (Phaseolus vulgaris L.) under different environments. BMC Genetics 13(1):1-9. 
Crossref

 

Padder BA, Sharma PN, Awale HE, Kelly JD (2017). Colletotrichum lindemuthianum, the causal agent of bean anthracnose. Journal of Plant Pathology 99(2):317-330.

View

 

Pallis E (2017). Big Data and Cloud Computing: A Survey of the State-of-the-Art and Research Challenges. In: Mavromoustakis C, Mastorakis G, Dobre C (eds). Advances in Mobile Cloud Computing and Big Data in the 5G Era. Studies in Big Data 22: 23-41. 
Crossref

 

Palloix A, Ayme V, Moury B (2009). Durability of plant major resistance genes to pathogens depends on the genetic background, experimental evidence and consequences for breeding strategies. New Phytologist 183(1):190-199. 
Crossref

 

Pamela P, Mawejje D, Ugen M (2014). Severity of angular leaf spot and rust diseases on common beans in Central Uganda. Uganda Journal of Agricultural Sciences 15(1):63-72.

 

Peña JM, Torres-Sánchez J, de Castro AI, Kelly M, López-Granados F (2013). Weed mapping in early-season maize fields using object-based analysis of unmanned aerial vehicle (UAV) images. PLoS One 8(10):e77151. 
Crossref

 

Persley D, Cooke T, House S (2010). Diseases of vegetable crops in Australia. CSIRO Publishing, Collingwood, Vic. P 292.
Crossref

 

Philipo M, Ndakidemi PA, Mbega ER (2020). Environmental and genotypes influence on seed iron and zinc levels of landraces and improved varieties of common bean (Phaseolus vulgaris L.) in Tanzania. Ecological Genetics and Genomics 15:100056. 
Crossref

 

Picon A, Seitz M, Alvarez-Gila A, Mohnke P, Ortiz-Barredo A, Echazarra J (2019). Crop conditional convolutional neural networks for massive multi-crop plant disease classification over cell phone acquired images taken on real field conditions. Computers and Electronics in Agriculture 167:105093.
Crossref

 

Rezende AA, Pacheco MTB, Silva VSND, Ferreira TAPDC (2018). Nutritional and protein quality of dry Brazilian beans (Phaseolus vulgaris L.). Food Science and Technology 38:421-427. 
Crossref

 

Said MK, Taher DI (2020). Characterization of partial resistance to bean rust disease caused by Uromyces appendiculatus in seven common bean varieties. Journal of Plant Pathology and Microbiology 11(7):506.

View

 

Shao FM, Teri JM (1981). Promising Phaseolus bean lines resistant to anthracnose in Tanzania. East African Agricultural and Forestry Journal 47(1-4):14-16. 
Crossref

 

Shrivastava S, Singh SK, Hooda DS (2017). Soybean plant foliar disease detection using image retrieval approaches. Multimedia Tools and Applications 76(24):26647-26674. 
Crossref

 

Shrivastava VK, Pradhan MK (2021). Rice plant disease classification using color features: A machine learning paradigm, Journal of Plant Pathology 103(1):17-26. 
Crossref

 

Shukla DD, Ward CW (1989). Identification and classification of potyviruses on the basis of coat protein sequence data and serology. Archives of Virology 106(3):171-200. 
Crossref

 

Singh RK, Tiwari A, Gupta RK (2022). Deep transfer modeling for classification of maize plant leaf disease. Multimedia Tools and Applications 81(5):6051-6067. 
Crossref

 

Singh V, Misra AK (2017). Detection of plant leaf diseases using image segmentation and soft computing techniques. Information Processing in Agriculture 4(1):41-49. 
Crossref

 

Skourletopoulos G, Mavromoustakis CX, Mastorakis G, Batalla JM, Dobre C, Panagiotakis S, Thresh JM (2003). Control of plant virus diseases in sub-Saharan Africa: the possibility and feasibility of an integrated approach. African Crop Science Journal 11(3):199-223. 
Crossref

 

Stavely JR, Pastor-Corrales MA (1989). Rust. p. 159-194.: HF Schwartz and MA Pastor-Corrales (ed.) Bean production problems in the tropics. Centro Internacional de Agricultura Tropical, Cali, Colombia. Rust. p. 159-194. In HF Schwartz and MA Pastor-Corrales (ed.) Bean production problems in the tropics. 2nd ed. Centro Internacional de Agricultura Tropical, Cali, Colombia.

 

Torres-Sanchez J, Lopez-Granados F, De Castro A, Pena-Barragan J (2013) Configuration and specifications of an unmanned aerial vehicle (UAV) for early site specific weed management. PLoS One. 8(3):e58210. 
Crossref

 

Tryphone GM, Chilagane LA, Nchimbi-Msolla S, Kusolwa PM (2015). Genetic characterization of angular leaf spot resistance in selected common bean landraces from Tanzania. African Journal of Biotechnology 14(43):2943-2948. 
Crossref

 

Tryphone GM, Chilagane LA, Protas D, Kusolwa PM, Nchimbi-Msolla S (2013). Marker assisted selection for common bean diseases improvements in Tanzania: prospects and future needs. Plant breeding from laboratories to fields. Intech 121-147.

 

Tugume JK, Tusiime G, Sekamate AM, Buruchara R, Mukankusi CM (2019). Diversity and interaction of common bacterial blight disease-causing bacteria (Xanthomonas spp.) with Phaseolus vulgaris L. The Crop Journal 7(1):1-7. 
Crossref

 

Vallejos CE, Astua-Monge G, Jones V, Plyler TR, Sakiyama NS, Mackenzie SA (2006). Genetic and molecular characterization of the I locus of Phaseolus vulgaris L. Genetics 172:1229-242.
Crossref

 

Verma S, Chug A, Singh A P (2020). Application of convolutional neural networks for evaluation of disease severity in tomato plant. Journal of Discrete Mathematical Sciences and Cryptography 23(1):273-282. 
Crossref

 

Wani AA, Bhat FA, Nisa RT, Wani TA, Raja W, Jan SK, Rather TR, Anjum N, Wani MA, Mohammad P (2022). Bean angular leaf spot (Phaeoisariopsis griseola): Epidemiology and yield loss assessment. The Pharma Innovation Journal 11(2):291-297.

 

Wasonga CJ, Pastor-Corrales MA, Porch TG, Griffiths PD (2010). Targeting gene combinations for broad-spectrum rust resistance in heat-tolerant snap beans developed for tropical environments. Journal of the American Society for Horticultural Science 135(6):521-532. 
Crossref

 

Wohleb CH (2011). Common Bacterial Blight and Halo Blight: Two Bacterial Diseases of Phytosanitary Significance for Bean Crops in Washington State. Washington State University Extension.

 

Worrall EA, Wamonje FO, Mukeshimana G, Harvey JJ, Carr JP, Mitter N (2015). Bean common mosaic virus and bean common mosaic necrosis virus: relationships, biology, and prospects for control. Advances in Virus Research 93:1-46. 
Crossref

 

Xiongkui H, Bonds J, Herbst A, Langenakens J (2017). Recent development for unmanned aerial vehicle for plant protection in East Asia. International Journal of Agricultural and Biological Engineering 10(3):18-30.

View

 

Yallappa D, Veerngouda M, Maski D, Palled V, Bheemanna M (2017). Development and evaluation of drone mounted sprayer for pesticide applications to crops. IEEE Global Humanitarian Technology Conference pp. 1-7. 
Crossref

 




          */?>