Wheat Rust Diseases Research Articles

Segmentation of Wheat Rust Disease Using Co-Salient Feature Extraction

Wheat Stripe Rust Disease (WRD) poses a significant threat to wheat crops, causing substantial yield losses and can result in total crop damage if not detected early. The localization of WRD-infected areas is a labor-intensive and time-consuming task due to the intricate and varied nature of the disease spread, especially for large plantations. Hence, segmentation of wheat crops is vital for early identification of the WRD-affected area, which allows for the implementation of targeted intervention measures. The state-of-the-art segmentation technique for WRD using the real-world semantic segmentation NWRD dataset is based on a UNet model with the Adaptive Patching with Feedback (APF) technique. However, this implementation is complex and requires significant resources and time for training due to the processing of each patch of the dataset. Our work in this paper improves the state-of-the-art by using a two-stage model: a Vision Transformer (ViT) classifier to distinguish between the rust and non-rust patches and a less complex co-salient object detection (Co-SOD) model for segmentation of the classified images. The Co-SOD model uses multiple rust patches to extract contextual features from a group of images. By analyzing multiple patches of wheat rust disease simultaneously, we can segment disease regions more accurately. Our results show that the proposed approach achieves a higher F1 score (0.638), precision (0.621), and recall (0.675) for the rust class with 5× less training time as compared to the previous works.

Phenotypic and Molecular Screening for Yellow Rust Resistance in Chinese Wheat Hybrids as Impacted by Selection

Exploring the potential of novel sources of wheat material is immensely important to combat the ever-changing pathogen causing wheat rust disease, associated with huge losses worldwide. This study focused on assessing rust (YR) resistance in Chinese wheat hybrids, combining molecular markers and field testing over two years with estimation of the impact of selection. Out of 416 hybrids, 108 were selected based on disease scoring parameters, showing negative genetic advancement for rust severity and coefficient of infection. The maximum rust severity reached 100% for four hybrids during the year 2017–18 whereas 70% for three hybrids during the year 2018–19. The maximum coefficient of infection was observed for four hybrids which were 100% (2017–18), whereas 70% for three hybrids (2018–19). Molecular genotyping with a set of five yellow rust resistance genes linked markers revealed that the most frequent molecular markers was Yr18/Lr34 (among the YR markers), present in 92 out of 108 hybrids. All the five Yr genes were present only in 18 hybrids. Yr18 was the most common resistance gene detected in 85% hybrids, while Yr17 was the least detected gene (57%). Notably, the average coefficient of infection (ACI) was not directly associated with the number of resistant genes. Considering the overall performance, the hybrid lines 17BH002, 17BH057, 17BH073, 17BH109, 17BH123, 17BH128, 17BH190, 18BH026 and 17BH155 had an overall low yellow rust severity and co-efficient of infection and could be considering promising lines for further utilization. Clustering based on rust resistance genes categorized hybrids into five main clusters for YR resistance genes, showing variability among these valuable genetic resources. The observed variability in this study can be utilized for future breeding programs focused on developing resistant wheat varieties.

Evaluation of Bread Wheat (Triticum aestivum L.) Germplasm for Wheat Rust Diseases Resistance and Yield Performance

The most important biotic constraints for wheat production in Ethiopia are Wheat rust diseases, insects, and weeds. Among wheat rust diseases, stem rust and yellow rust are the more limiting factors for yield. A study was undertaken at Kulumsa and Melkasa in alpha lattice design with two replications from July to December 2023 to evaluate the performance of introduced bread wheat materials from CIMMYT for wheat rust diseases and yield. Data were collected for stem rust, yellow rust, and agronomic traits, including grain yield. The ANOVA showed high and significant genotypic variation (p<0.05) on Days to heading (DTH), Hectoliter weight (HLW), Thousand kernels weight (TKW), and Yield (YLD). Then, the result from LSD mean comparision revealed genotypes: EBW222102 YLD=7.7 t ha-1, EBW222106 YLD= 7.61 t ha-1, EBW222108 YLD=7.66 tha-1, EBW222136 YLD= 7.76 t ha-1 to be significantly higher than the check variety Daka YLD= 6.22 t ha-1 (p<0.05) at Kulumsa. Genotypes: EBW222108, EBW222109, EBW222111, EBW222112, EBW222114, EBW222125, EBW222142, and EBW222143 showed resistance to moderate resistance for yellow rust and stem rust across both locations. About eighteen genotypes were susceptible to stem rust diseases with the coefficient of infection YRCI≥30 at Melkasa. Genotypes EBW222111, EBW222126, EBW222129, EBW222134, and EBW222142 delivered high yield, greater than or equal to 2 t ha-1 at Melkasa. Generally, selection for disease resistance and high-yielding genotypes enable the development and release of noble varieties for wheat-producing farmers.

Wheat Rust Disease Detection Using Convolutional Neural Network

Wheat is a vital crop globally, playing a crucial role in ensuring food security. Any damage to wheat crops, particularly from diseases like wheat rust, can exacerbate the global food crisis. Wheat rust, along with other crop diseases, poses a serious threat to agricultural sustainability and food supply. Technological approaches to detect crop health can significantly enhance disease prevention compared to traditional manual methods. This study focuses on detecting wheat rust diseases, specifically wheat leaf rust, stem rust, and yellow rust, all of which vary in their impact on crops. By employing image analysis of wheat plants, the research aims to accurately distinguish between healthy crops and those affected by rust diseases. The experiments consider multiple variables, such as learning rate, dropout, and train-test split ratio, reflecting a comprehensive research approach. The model achieved an impressive 99.64% accuracy in detecting wheat rust, highlighting its potential for early disease detection and prevention.

Phenotypic Selection of Bread Wheat (Triticum aestivum L.) Elite Lines for Diseases Resistance and Yield Potential

Wheat diseases and pests globally threaten wheat production. Wheat rust diseases are the most distractive biotic constraint of wheat in Ethiopia. Therefore, this experiment aimed to select and develop wheat rust resistance bread wheat varieties for the farmers. The result of the study revealed highly significant variation existed among tested genotypes for Plant Height (PHT), Days to Heading (DTH), Days to Maturity (DTM), Thousand Kernel Weight (TKW), Hectoliter Weight (HLW), and Yield (YLD) at (P< 0.001). Out the forty-nine introduced genotypes, five genotypes: EBW222153, EBW222159, EBW222160, EBW222162, and EBW222164, showed resistance to moderately resistance for both yellow rust and stem rust diseases across the two test locations. Also, they delivered high average yield: 5.9 tha-1, 4.46 tha-1, 5.17tha-1,4.15 tha-1, and 4.79tha-1 respectively. Furthermore, all the above-mentioned genotypes, except EBW222162, showed significantly higher yield at (P<0.05) than the check variety, Shaki. Furthermore, EBW222153, EBW222159, EBW222160, EBW222162, EBW222164, and EBW222178 would resistance to moderately resistance for stem rust at Melkasa. Thus, the selection of these genotypes and advancing to the next stage of breeding pipelines enables to release noble bread wheat varieties for the farmers.

5-mC methylation study of sORFs in 3'UTR of transcription factor JUNGBRUNNEN 1-like during leaf rust pathogenesis in wheat.

JUB1, a NAC domain containing hydrogen peroxide-induced transcription factor, plays a critical role in plant immunity. Little is known about how JUB1 responds to leaf rust disease in wheat. Recent discoveries in genomics have also unveiled a multitude of sORFs often assumed to be non-functional, to argue for the necessity of including them as potential regulatory players of translation. However, whether methylation on sORFs spanning the 3'UTR of regulatory genes like JUB1 modulate gene expression, remains unclear. In this study, we identified the methylation states of two sORFs in 3'UTR of a homologous gene of JUB1 in wheat, TaJUB1-L, at cytosine residues in CpG, CHH and CHG sites at different time points of disease progression in two near-isogenic lines of wheat (HD2329), with and without Lr24 gene during leaf rust pathogenesis. Here, we report a significant demethylation of the CpG dinucleotides occurring in the sORFs of the 3'UTR in the resistant isolines after 24h post-infection. Also, the up-regulated gene expression observed through RT-qPCR was directly proportional to the demethylation of the CpG sites in the sORFs. Our findings indicate that TaJUB1-L might be a positive regulator in providing tolerance during leaf rust pathogenesis and cytosine methylation at 3'UTR might act as a switch for its expression control. These results enrich the potential benefit of conventional methylation assay techniques for unraveling the unexplored enigma in epigenetics during plant-pathogen interaction in a cost-effective and confidentially conclusive manner.

Identification of hub genes associated with stripe rust disease in wheat through integrative transcriptome and gene-based association study

Stripe rust poses a severe menace to wheat cultivation worldwide, occurring as the most devastating disease threatening global wheat industry. The use of resistance genes has been considered as an effective method for management of rust diseases. However, for most of yellow rust (Yr) genes, the molecular mechanisms and pathways responsible for resistance response against Puccinia striiformis (Pst) infection are not well understood. In this study, transcriptome analysis was conducted in two cultivars viz. PBW343 (susceptible) and FLW30 (resistant) to the pathotype 46S119 of Pst in India. The inoculation with Pst pathotype 46S119 was followed by leaf sample collection at different hours per inoculation (hpi) viz.12, 48, and 72. The transcriptome analysis revealed a total of 18,482 differentially expressed genes (DEGs) across different time points. However, a set of 159 genes were identified to exhibit a genome-wide association with P ≤ 0.05 level of significance, which can be represented as Nominally Significant GWAS Geneset (NSGG). By utilizing Weighted Gene Co-expression Network Analysis (WGCNA), we identified 924 and 79 genes from the DEGs and NSGG dataset belonging to different modules. Functional validation of genes was done to confirm the transcriptional expression of four potential genes, including two hub genes associated with the defence mechanism against rust infection.

Unveiling wheat rust epidemics: A long‐term surveillance study in Ethiopia's North Western Amhara region

AbstractWheat is one of the most significant food security crops in Ethiopia. However, the production and productivity of wheat is far below its potential due to many challenges. Wheat disease epidemics, typified by wheat rusts, are among the most important challenges, causing large‐scale production losses that put farmers’ income and food security at risk. We analysed and presented the long‐term and comprehensive wheat rust surveillance results to better understand the past rust disease occurrences, ultimately contributing to future wheat production and rust control strategies. A total of 1140 wheat fields have been surveyed from 2009 to 2021 throughout four administrative zones of the North Western Amhara region. We identified spatial and temporal trends in the distribution and status of wheat stem rust (Sr), yellow rust (Yr) and leaf rust (Lr) using a combination of different statistical tools and data‐analysis methods. The results of the study indicated that 52.5%, 11.3% and 3.4% of the wheat fields were infected with yellow, stem and leaf rust, respectively. Even so, the distribution and intensity of wheat rust disease vary across locations, elevations and cultivated varieties. The associations between altitude and disease prevalence show a linear correlation. We also identified hotspots of yellow and stem rust, which will be used by the country's wheat rust disease early warning system. In order to reverse low varietal diversification in the farmer's field, it is necessary to work with all stakeholders involved in the wheat production sector in a better synergy.

Assessing leaf rust resistance in Pakistani wheat landraces and its impact on grain yield

Leaf rust, caused by the fungus Puccinia triticina, is the most widespread rust disease in wheat, involving two hosts: common wheat as the primary host and specific plants like Thalictrum speciosissimum or Isopyrum fumaroides as alternate hosts. The present study investigated the crucial interplay between leaf rust resistance and grain yield in 10 Pakistani wheat landraces. By employing comprehensive assessments, including field trials and molecular analyses, the research sheds light on the resistance mechanisms within these landraces. Moreover, it examines the direct correlation between leaf rust resistance levels and the resultant grain yield, unraveling critical insights into the potential enhancement of wheat cultivars amidst the escalating challenges posed by this devastating disease. The findings hold promise for optimizing agricultural strategies, ensuring sustainable wheat production, and safeguarding global food security. The study focused on 10 bread wheat landraces, examining traits crucial for high-yielding varieties. Traits such as days to heading, days to maturity, grain filling period, flag leaf area, peduncle length, spike length, grains per spike, as well as measurements like 10-grain width, 10-grain length, and 100-grain weight were recorded. The analysis of variance (ANOVA) for these ten genotypes highlighted significant differences (P0.05 and P ≤ 0.01) across all traits.

Management of Rust in Wheat Using IPM Principles and Alternative Products

Overall, there is a major wish that European farmers implement integrated pest management (IPM), particularly to reduce dependence on pesticides. In the European Rustwatch project, partners conducted nineteen trials across nine different countries during 2020 and 2021 to investigate different IPM strategies, focusing on controlling rust diseases in winter wheat. The trials included the use of varieties with contrasting levels of resistance, variety mixtures, reduced fungicide rates, thresholds, and Decision Support Systems (DSSs), and testing alternative products to fungicides. Sixteen trials developed yellow rust (Puccinia striiformis f. sp. tritici) infections, and six trials developed brown rust (Puccinia triticina) infections. Resistant varieties proved highly effective in keeping down yellow rust infection, and variety mixtures also effectively reduced infection levels and stabilized yields. Rust was fully controlled using 25% of standard fungicide rates, even under high disease pressure. Using DSSs provided sufficient control of rust diseases and resulted in competitive net economic returns due to fewer fungicide applications. The alternative products tested included two biological control agents and four alternative chemistries, which all gave inferior and insufficient control against rust compared with chemical fungicides. The trial work demonstrated that there are good and reliable options for including IPM into disease control in wheat.

Response of Durum Wheat (Triticum durum L.) Genotypes to Blended NPSB Fertilizer Rates and Wheat Rust Diseases in Southern Ethiopia

Lack of improved varieties and soil fertility degradation are among the major factors that limit durum wheat (Triticum turgidum L.) production in Ethiopia. The objectives of this study were to determine the response of durum wheat cultivars (i) to nitrogen, phosphorus, sulphur, and boron (NPSB) fertilizer blend; (ii) rust infection; and (iii) partial budget analysis using NPSB fertilizer blend. A 2-location experiment was conducted in 2021 in South Ethiopia to evaluate the response of 4 durum wheat genotypes to blended NPSB fertilizer rates in a randomized complete block design in a 4 × 4 factorial arrangement. The results showed significant (P<0.05) differences due to genotypes (G), rate of NPSB fertilizer (F), and location (E) for all the parameters. The effects due to the G × F × E interaction were significant (P<0.05) for the number of total tillers, number of productive tillers, and grain yield (GY). A significant (P<0.05) G × E interaction was observed for the majority of parameters except days to 50% heading and plant height. Durum wheat genotype S3-213037 produced the highest grain yield of 3689.3 kg·ha−1 when NPSB fertilizer was applied at a rate of 225 kg·ha−1 at Adekoysha. Moreover, partial budget analysis indicated that genotype S3-213037 accrued a maximum benefit of 112,986.50 birr·ha−1 and the highest marginal rate of return of 1356.4% when NPSB fertilizer was applied at a rate of 225 kg·ha−1. Therefore, this study indicated that the application of an NPSB fertilizer rate of 225 kg·ha−1 could be used in the production of durum wheat in the East African region with similar agroecologies to the study area.

The F-box protein encoding genes of the leaf-rust fungi Puccinia triticina: genome-wide identification, characterization and expression dynamics during pathogenesis.

The F-box proteins in fungi perform diverse functions including regulation of cell cycle, circadian clock, development, signal transduction and nutrient sensing. Genome-wide analysis revealed 10 F-box genes in Puccinia triticina, the causal organism for the leaf rust disease in wheat and were characterized using in silico approaches for revealing phylogenetic relationships, gene structures, gene ontology, protein properties, sequence analysis and gene expression studies. Domain analysis predicted functional domains like WD40 and LRR at C-terminus along with the obvious presence of F-box motif in N-terminus. MSA showed amino acid replacements, which might be due to nucleotide substitution during replication. Phylogenetic analysis revealed the F-box proteins with similar domains to be clustered together while some sequences were spread out in different clades, which might be due to functional diversity. The clustering of Puccinia triticina GG705409 with Triticum aestivum TaAFB4/TaAFB5 in a single clade suggested the possibilities of horizontal gene transfer during the coevolution of P. triticina and wheat. Gene ontological annotation categorized them into three classes and were functionally involved in protein degradation through the protein ubiquitination pathway. Protein-protein interaction network revealed F-box proteins to interact with other components of the SCF complex involved in protein ubiquitination. Relative expression analysis of five F-box genes in a time course experiment denoted their involvement in leaf rust susceptible wheat plants. This study provides information on structure elucidation of F-box proteins of a basidiomycetes plant pathogenic fungi and their role during pathogenesis.

A novel hybrid segmentation technique for identification of wheat rust diseases

A novel hybrid segmentation technique for identification of wheat rust diseases

Stripe Rust Effector Pst_9302 Inhibits Wheat Immunity to Promote Susceptibility.

Puccinia striiformis f. sp. tritici is an obligate biotrophic fungus that causes destructive stripe rust disease in wheat. During infection, Pst secretes virulence effectors via a specific infection structure-the haustorium-inside host cells to disturb host immunity and promote fungal colonization and expansion. Hence, the identification and functional analyses of Pst effectors are of great significance in deciphering the Pst pathogenicity mechanism. Here, we identified one candidate Pst effector Pst_9302 that could suppress Bax-triggered cell death in Nicotiana benthamiana. qRT-PCR analyses showed that the transcript levels of Pst_9302 were highly increased during the early infection stages of Pst. The transient expression of Pst_9302 in wheat via the type-three secretion system (T3SS) significantly inhibited the callose deposition induced by Pseudomonas syringae EtHAn. During wheat-Pst interaction, Pst_9302 overexpression suppressed reactive oxygen species (ROS) accumulation and cell death caused by the avirulent Pst race CYR23. The host-induced gene silencing (HIGS) of Pst_9302 resulted in decreased Pst pathogenicity with reduced infection area. The results suggest that Pst_9302 plays a virulence role in suppressing plant immunity and promoting Pst pathogenicity. Moreover, wheat voltage-dependent anion channel 1 protein (TaVDAC1) was identified as candidate Pst_9302-interacting proteins by yeast two-hybrid (Y2H) screening. Pull-down assays using the His-Pst_9302 and GST-TaVDAC1 protein verified their interactions. These results suggest that Pst_9302 may modulate wheat TaVDAC1 to regulate plant immunity.

Identification and Categorization of Yellow Rust Infection in Wheat through Deep Learning Techniques

The global wheat industry faces significant challenges due to yellow rust disease, This is induced by fungus Puccinia striiformis, as it leads to substantial crop losses and economic impacts. Timely detection and classification of the disease are essential for its effective management and control. In this study, we investigate the potential of DL and ML techniques for detecting and classifying yellow rust disease in wheat. We utilize three state-of-the-art CNN models, namely ResNet50, DenseNet121, and VGG19, to analyze wheat leaf images and extract relevant features. These models were developed and refined using a large dataset of annotated wheat photos. Encompassing both healthy plants and those affected by yellow rust disease. Furthermore, we examine the effectiveness of data augmentation and transfer learning in enhancing classification performance. Our findings reveal that the DL-based CNN models surpass traditional machine learning techniques in detecting and classifying yellow rust disease in wheat. Among the tested CNN models, EfficientNetB3 demonstrates the best performance, emphasizing its suitability for large-scale and real-time monitoring of wheat crops. This research contributes to the development of precision agriculture tools, laying the groundwork for prompt intervention and management of yellow rust disease, ultimately minimizing yield loss and economic impact on wheat production.

The potential of UAV and very high-resolution satellite imagery for yellow and stem rust detection and phenotyping in Ethiopia

Very high (spatial and temporal) resolution satellite (VHRS) and high-resolution unmanned aerial vehicle (UAV) imagery provides the opportunity to develop new crop disease detection methods at early growth stages with utility for early warning systems. The capability of multispectral UAV, SkySat and Pleiades imagery as a high throughput phenotyping (HTP) and rapid disease detection tool for wheat rusts is assessed. In a randomized trial with and without fungicide control, six bread wheat varieties with differing rust resistance were monitored using UAV and VHRS. In total, 18 spectral features served as predictors for stem and yellow rust disease progression and associated yield loss. Several spectral features demonstrated strong predictive power for the detection of combined wheat rust diseases and the estimation of varieties’ response to disease stress and grain yield. Visible spectral (VIS) bands (Green, Red) were more useful at booting, shifting to VIS–NIR (near-infrared) vegetation indices (e.g., NDVI, RVI) at heading. The top-performing spectral features for disease progression and grain yield were the Red band and UAV-derived RVI and NDVI. Our findings provide valuable insight into the upscaling capability of multispectral sensors for disease detection, demonstrating the possibility of upscaling disease detection from plot to regional scales at early growth stages.

Candidate effectors for leaf rust resistance gene Lr28 identified through transcriptome and in-silico analysis.

Puccinia spp. causing rust diseases in wheat and other cereals secrete several specialized effector proteins into host cells. Characterization of these proteins and their interaction with host's R proteins could greatly help to limit crop losses due to diseases. Prediction of effector proteins by combining the transcriptome analysis and multiple in-silico approaches is gaining importance in revealing the pathogenic mechanism. The present study involved identification of 13 Puccinia triticina (Pt) coding sequences (CDSs), through transcriptome analysis, that were differentially expressed during wheat-leaf rust interaction; and prediction of their effector like features using different in-silico tools. NCBI-BLAST and pathogen-host interaction BLAST (PHI-BLAST) tools were used to annotate and classify these sequences based on their most closely matched counterpart in both the databases. Homology between CDSs and the annotated sequences in the NCBI database ranged from 79 to 94% and with putative effectors of other plant pathogens in PHI-BLAST from 24.46 to 54.35%. Nine of the 13 CDSs had effector-like features according to EffectorP 3.0 (≥0.546 probability of these sequences to be effector). The qRT-PCR expression analysis revealed that the relative expression of all CDSs in compatible interaction (HD2329) was maximum at 11 days post inoculation (dpi) and that in incompatible interactions (HD2329 + Lr28) was maximum at 3 dpi in seven and 9 dpi in five CDSs. These results suggest that six CDSs (>0.8 effector probability as per EffectorP 3.0) could be considered as putative Pt effectors. The molecular docking and MD simulation analysis of these six CDSs suggested that candidate Lr28 protein binds more strongly to candidate effector c14094_g1_i1 to form more stable complex than the remaining five. Further functional characterization of these six candidate effectors should prove useful for a better understanding of wheat-leaf rust interaction. In turn, this should facilitate effector-based leaf rust resistance breeding in wheat.

In-silico prediction, characterization, molecular docking and dynamic simulation studies for screening potential fungicides against leaf rust of Triticum aestivum

Triticum aestivum is an important crop worldwide, which is a large source of food grain. T.aestivum demands on developed countries will grow every year, this increase in the demand is profoundly serious especially in the light climate change which would lead to a 29% reduction in final productivity. Rust fungus attacks the T.aestivum, specifically newly planted T.aestivum plants, which block the vascular system, stun, and finally damage grain and tillers. In present study we predict the 3D structure then find the binding pocket and conserved domains for MAPkinase-1 of Puccinia triticina. After that, screen the FungiPAD, PubChem, NPAtlas databases by physicochemical properties, docking, clustering, ADME (Absorption, distribution, metabolism, and excretion) and PAINS (pan assay interference compounds) filter analysis. Through this screening process screen the nine compounds, which are benzovindiflupyr, furametpyr, isopyrazam, fenaminstrobin, and flumorph from Fungicide database: zoxamide, vinclozolin, pentachloronitrobenzene, and dithianon from PubChem database, based on the binding energy, clustering, ADME and PAINS analysis. All these nine compounds bind in the same pocket and show the same pattern of interaction. Among these nine compounds, select the two compounds (PubChem:122087 (-6.96 kcal/mol) and FDBD02904 (-8.62 kcal/mol)) based on binding energy for 100 ns MD simulation and free energy calculation. MD simulation shows stability throughout the simulation, and it shows the sable interaction when compounds bind to the MAPKinase 1 protein which may help to protein kinase pathways in plant defense response. This result helps to design alternative fungicide against the wheat rust disease. Communicated by Ramaswamy H. Sarma

Integrated digital image processing techniques and deep learning approaches for wheat stripe rust disease detection and grading

This study proposes integrated image processing and deep learning approaches for wheat rust disease detection and grading. Wheat stripe rust is one of the diseases affecting the crop in nearly all wheat-growing countries, where humidity and high temperatures characterize the growing season. The early and prompt detection of this disease is essential for establishing timely remedial measures to contain its spread. We present several image-processing and deep-learning approaches for detecting wheat stripe rust. The image processing techniques used in this research include single-band and dual-band processing, computing the Visual Atmospheric Resistance Index (VARI), image segmentation, and image cropping. These techniques are evaluated with different deep learning approaches like transfer learning and other Convolutional Neural Networks (CNNs) architectures. The results obtained on the dataset show that processing the images to eliminate the background as much as possible before training a suitable deep learning model with a low number of epochs optimally classifies the wheat stripe rust into the target classes (i.e., healthy, resistant, and susceptible) with an overall accuracy of 84.1% on the holdout test set. Besides the comparative analysis of various image processing approaches with different deep learning models, this study suggests multiple future research directions.

The NWRD Dataset: An Open-Source Annotated Segmentation Dataset of Diseased Wheat Crop.

Wheat stripe rust disease (WRD) is extremely detrimental to wheat crop health, and it severely affects the crop yield, increasing the risk of food insecurity. Manual inspection by trained personnel is carried out to inspect the disease spread and extent of damage to wheat fields. However, this is quite inefficient, time-consuming, and laborious, owing to the large area of wheat plantations. Artificial intelligence (AI) and deep learning (DL) offer efficient and accurate solutions to such real-world problems. By analyzing large amounts of data, AI algorithms can identify patterns that are difficult for humans to detect, enabling early disease detection and prevention. However, deep learning models are data-driven, and scarcity of data related to specific crop diseases is one major hindrance in developing models. To overcome this limitation, in this work, we introduce an annotated real-world semantic segmentation dataset named the NUST Wheat Rust Disease (NWRD) dataset. Multileaf images from wheat fields under various illumination conditions with complex backgrounds were collected, preprocessed, and manually annotated to construct a segmentation dataset specific to wheat stripe rust disease. Classification of WRD into different types and categories is a task that has been solved in the literature; however, semantic segmentation of wheat crops to identify the specific areas of plants and leaves affected by the disease remains a challenge. For this reason, in this work, we target semantic segmentation of WRD to estimate the extent of disease spread in wheat fields. Sections of fields where the disease is prevalent need to be segmented to ensure that the sick plants are quarantined and remedial actions are taken. This will consequently limit the use of harmful fungicides only on the targeted disease area instead of the majority of wheat fields, promoting environmentally friendly and sustainable farming solutions. Owing to the complexity of the proposed NWRD segmentation dataset, in our experiments, promising results were obtained using the UNet semantic segmentation model and the proposed adaptive patching with feedback (APF) technique, which produced a precision of 0.506, recall of 0.624, and F1 score of 0.557 for the rust class.