Background: Faba bean is an important legume crop valued for its nutritional and soil-enriching benefits, yet its productivity is severely affected by foliar diseases. Automated image-based detection using deep learning provides a rapid and reliable approach for early disease identification and improved crop management. Methods: This study developed a machine learning-based automated framework for multi-class classification of Faba bean leaf diseases using transfer learning with the VGG16 convolutional neural network. A dataset of 8,021 RGB images collected under natural field conditions was used, comprising four classes: healthy, rust, gall and chocolate spot. Images were resized to 224 × 224 pixels and normalized prior to training. The pretrained convolutional layers of VGG16 were frozen and a custom classification head with global average pooling and dropout regularization was added. Model performance was evaluated using classification metrics. Result: The proposed model achieved an overall classification accuracy of 92.34% and a macro-averaged F1-score of 0.9227 on the test dataset. Strong classification performance was observed across all disease categories, with particularly high predictive accuracy for healthy and rust classes. The findings demonstrate the effectiveness of transfer learning for plant disease detection and highlight its potential for scalable, automated crop health monitoring in precision agriculture.
Faba bean yield reflects complex relationships among genotype, environment, and agronomic traits. This study evaluated an explainable Tabular Prior-data Fitted Network (TabPFN) framework for estimating plot-mean single-plant yield and prioritizing traits using 398 plot-level observations, 13 measured agronomic predictors, and six derived features. On the reference 80/20 split, TabPFN achieved the best values for all four test metrics (R2 = 0.8746, RMSE = 1.9132 g plant−1, MAE = 1.0819 g plant−1, and MAPE = 8.16%). The Friedman test detected differences among the six models (χ2(5) = 16.75, p = 0.005); Nemenyi comparisons distinguished TabPFN from HistGradientBoosting and SVR, whereas the Holm-corrected Wilcoxon analysis confirmed only the TabPFN–SVR difference. Across 10 repeated 80/20 splits, TabPFN obtained the highest mean test R2 (0.8614 ± 0.0691), ranked first in eight splits, and produced a higher R2 than every tuned baseline in at least eight splits. SHAP, permutation importance, and LOCO analyses emphasized pod-, seed-, and biomass-related predictors. Repeated-split ablation showed that derived features improved TabPFN consistently, whereas removing selected target-proximal yield variables reduced performance for every model. The framework is therefore a harvest-time trait-estimation and trait-prioritization tool rather than an early-season forecasting system. Notably, TabPFN achieved this performance without the 100-trial Optuna search used for each baseline; only n_estimators was screened over four prespecified values.
The precise identification of microspore and pollen at the optimal developmental stages to be induced towards embryogenesis (vacuolated microspores and young pollen) is essential for induction of in vitro androgenesis in plants. Such identification is not always easy, and it is especially difficult in recalcitrant species such as Vicia faba . The present study evaluates the relationship between floral bud and anther morphometric parameters, and microspore/pollen developmental stages in V. faba using various methods including fluorescent staining, differential interference contrast microscopy, and morphometry. We measured flower bud and anther length and width, grouping them at different intervals, and performed a detailed microscopical and anatomical analysis of buds, anthers and microspores/pollen at different stages. Our results demonstrated that flower buds in V. faba exhibit complex and irregular morphologies, with considerable variation in both sepal length and shape. Furthermore, the determination of microspore and pollen developmental stages in this species is constrained by pronounced developmental asynchrony and strong genotype dependence. Although anther length measurements correlate closely with microspore and pollen developmental stages, their practical use can be challenging. Therefore, measuring flower bud length, while excluding sepals, remains the most practical criterion for routine applications. Combining this refined morphometric approach with microscopic validation appears to be the most effective strategy for improving the identification of flower buds containing microspores or pollen at developmental stages suitable for androgenesis induction in this recalcitrant legume species.
Abstract Multispectral three‐dimensional (3D) imaging offers substantial potential for plant phenotyping, yet manual segmentation of plant organs remains a bottleneck in breeding programs. We developed a color‐based filtering workflow for faba bean ( Vicia faba L.) point clouds that optimizes lower and upper thresholds of spectral indices and broadband brightness via Bayesian optimization. Rather than maximizing geometric segmentation accuracy, thresholds are selected to maximize correlations between retained points and yield‐related traits, outperforming manual filtering, reducing user effort, and standardizing decisions. Across multispectral 3D point clouds, Bayesian optimization recovered index‐specific threshold ranges that yielded strong in‐sample correlations with grain yield ( r = 0.72), bean number ( r = 0.61), pod number ( r = 0.53), and straw biomass ( r = 0.72). Peak associations occurred at harvest for straw biomass, at 41 days before harvest (DBH) for seed yield, 33 DBH for bean number, and 34 DBH for pod number. Across the season, greenness‐based indices and broadband brightness metrics consistently showed stronger links with seed yield than pigment ratio or water status indices. For straw biomass and pod number, pigment ratio indices showed consistently lower correlations. Targeting trait‐relevant canopy signals via Bayesian optimization enables reliable, nondestructive assessment of relationships between spectral signals and yield‐related traits in faba bean. By optimizing thresholds to maximize trait correlations rather than geometric accuracy, the workflow can support earlier, more cost‐efficient identification of high‐performing genotypes under drought stress and contribute to strengthening high‐throughput phenotyping in breeding programs. This enables faster identification of canopy signals most relevant to target traits.
Abstract Faba bean ( Vicia faba L.) is a widely cultivated legume in temperate regions, valued for human consumption, animal feed, hay production, and as a cover crop. Improving faba bean productivity requires accurate characterization of morphological and physiological traits that govern crop performance and yield. Although conventional phenotyping methods are available, they are often constrained by high cost, limited accuracy, and insufficient spatial and temporal coverage. In recent years, high-throughput phenotyping (HTP) approaches have shown potential to overcome these limitations. HTP integrates sensors, unoccupied aerial and ground vehicles, and imaging systems to enable rapid, and non-destructive monitoring of crop traits at high spatiotemporal resolution. Despite increasing adoption of HTP, no study has comprehensively compared and synthesized these methods for faba beans, therefore is the goal of this study with emphasis on advanced sensing and data analytics including machine learning (ML) and deep learning (DL). A systematic evaluation of 24 peer-reviewed articles from an initial pool of 381 publications between 2015 and 2025, identified research trends, performance benchmarks, and integration challenges with HTP-based faba bean characterization. Substantial increase in faba bean HTP studies has been noted after 2021, with ML approaches dominating current applications (37.5%). Most studies have relied on small datasets, single-season experiments, and limited environmental variability, restricting model robustness and scalability. Such limitations, research gaps, and future directions are also outlined to support reliable, and scalable phenotyping for improved faba bean production.
The apoplast of leaves is involved in nutrient transport, microbe-host interactions, systemic signaling, cell wall dynamics, and serves as an interface for various other physiological processes. The composition of the apoplastic solute pool, which supports many of these functions, is dynamic and shaped by developmental and environmental cues. However, due to the complexity and compartmentalization of the apoplast, analysing these fluids - and thus the associated physiological processes - remains technically challenging. This study introduces a minimally invasive method for extracting apoplastic fluids from leaves of selected dicots (e.g. Arabidopsis thaliana, Vicia faba, and many more), offering two key advantages: (i) repeated extractions from the same leaves to generate time-series data, such as every 24 hours, over consecutive days, and (ii) high spatial resolution, enabling identification of macrodomains within the leaf apoplast. For example, abscisic acid macrodomains were revealed along the leaf axis, providing insight into apoplastic hormone regulation. The method also reveals other previously unrecognized aspects, such as the accumulation of kaempferol glycosides in the apoplast after plants experienced salt stress. Finally, the method addresses the distortion of apoplast compound levels caused by dilution bias that results from the inconsistent recovery of infiltration fluid. Adding pyranine enables correction, ensuring accurate and comparable data. By integrating spatial and temporal precision, this new tool will promote a deeper understanding of plant apoplastic processes and their physiological relevance in various biological contexts.
Sylvain Poque · Ulrika Carlson-Nilsson · Muhammad Omer · Anna Palmé · Ingunn M. Vågen · G. Poulsen · Matti W. Leino · Kristiina Himanen · Hamid Khazaeı
Abstract Faba bean ( Vicia faba L.) has great potential to contribute to sustainable agriculture and protein security globally but is known to be very sensitive to drought stress. Uncovering drought-adapted germplasm is critical for developing resilient cultivars and advancing our understanding of the mechanisms underlying stress adaptation. However, high-throughput plant phenotyping under stress conditions remain a major bottleneck in crop genetics and breeding programs. In this study, a multi-sensor indoor phenotyping platform was used to assess 44 faba bean genotypes under water deficit conditions. Standardized, monitored stress conditions were achieved by watering-by-weighing for drought onset, duration, and intensities allowing genotype-level comparisons. The genotypes showed a range of stress responses in growth and physiology, including traits such as plant height, biomass, water use efficiency (WUE), and chlorophyll fluorescence parameters. Digital biomass, derived from combined top- and side-view plant imaging, was strongly correlated with biological biomass at the experimental endpoint, validating its use as a non-destructive proxy for growth assessment in faba bean. Time-resolved generalized additive modelling further revealed genotype-specific differences in the timing and magnitude of water deficit response. Genotypes that maintained growth and WUE under water deficit conditions may serve as valuable pre-breeding materials for development of drought-adapted faba bean.
This study established an integrated analytical method based on near-infrared spectroscopy (NIRS) for the rapid, non-destructive, and quantitative detection of four major nutritional components in faba beans: starch, protein, moisture, and dietary fiber. By systematically comparing individual and combined spectral preprocessing strategies, optimal preprocessing combinations for each component were identified. Seven feature wavelength selection algorithms, including Competitive Adaptive Reweighted Sampling (CARS), were employed to extract key spectral variables. Predictive models were subsequently developed using four modeling approaches: Partial Least Squares (PLS), Random Forest (RF), Support Vector Machine (SVM), and Multilayer Perceptron (MLP). The results demonstrated that combined preprocessing methods significantly outperformed single techniques. The CARS algorithm exhibited the most robust performance in feature extraction, and the MLP model consistently surpassed traditional machine learning methods in predicting all components. The optimal modeling pipelines for each component were ultimately determined as follows: starch (MLP + CARS + MSC + SG + MSS, R 2 = 0.92), protein (MLP + CARS + SD + SNV + MSC + MSS, R 2 = 0.94), moisture (MLP + SPA + SG + SNV, R 2 = 0.9973), and dietary fiber (MLP + PCA + FD + SNV, R 2 = 0.9999). This study verifies the effectiveness of combining NIRS with deep learning for the simultaneous detection of multiple components in faba beans and provides a reliable methodological framework for the non-destructive quality assessment of agricultural products.
Though currently a minor crop, faba bean is a promising source of plant-based protein as global diets shift towards more plant-based nutrition. To realise this potential, advances in breeding and cultivation are crucial. To exploit heterosis, faba bean breeding frequently utilises synthetic cultivars, which involves open pollination of inbred lines to produce a mixture of F 1 hybrid seeds and self-pollinated offspring. Pure F 1 hybrid cultivars are currently unavailable due to unstable cytoplasmic male sterility (CMS) systems. An ability to distinguish F 1 seeds from their parental inbreds via characteristics associated with xenia effects could change this. The xenia effect refers to the influence of paternal pollen on seed traits, for example seed weight and cotyledon cells in faba bean. In this study, we exploited the xenia effect captured in hyperspectral imaging data to develop machine learning scenarios for discriminating between parental and F 1 seeds of open pollinated synthetic combinations (Syn-1). The hyperspectral data were pre-processed using Savitzky–Golay filtering to reduce noise and smooth the spectra. Various machine learning algorithms were applied, incorporating Bayesian hyperparameter optimisation. The scenarios achieved up to 98.9 % accuracy in separating parental components of Syn-1. When including all seeds, the model achieved 40.7 %, indicating moderate detection and classification performance. As the harmonic mean of precision and recall, the F1 score accounts for both the correctness of F 1 seed detections and the completeness with which F 1 seeds were detected. While this approach does not yet enable the development of full hybrid cultivars, it paves the way for hybrid-enriched cultivars. These could help to streamline breeding for synthetic cultivars and potentially increase yields, for example by increasing the proportion of F 1 hybrid seeds in synthetic cultivars. This study extends knowledge of the xenia effect in faba bean and provides a basis for further research aimed at enhancing breeding methods and productivity.
Abstract Background Faba bean is an important grain legume in temperate cropping systems because it provides protein-rich seed and contributes biological nitrogen fixation. However, its productivity is highly sensitive to drought, and breeding for improved drought performance is constrained by complex genotype by environment interactions and the difficulty of measuring relevant traits at scale. This study evaluated whether scanner-derived vegetation indices (VI), 3D canopy traits, and their combination can predict key agronomic and physiological traits in drought-stressed faba bean, and how predictive ability changes when information is used from single dates or cumulatively across the season. Results Predictive performance was strongly trait dependent and varied with predictor set and temporal strategy. Combined VI + 3D predictors generally produced the highest and most consistent predictive ability for major traits. Total grain yield reached 0.75 under cumulative VI + 3D prediction at 93 days after sowing (DAS 93), cumulative water uptake peaked at 0.80 at DAS 97, and total straw biomass reached 0.66 at DAS 104. In contrast, some component traits were predicted equally well or better by 3D information alone, including grain number with 0.70 and pod number with 0.55 under cumulative 3D prediction. Useful prediction windows also differed among traits, with broad late-season windows for major agronomic traits but narrower, more stage-specific windows for productive tillers, thousand kernel weight, and water-use efficiency. Conclusion Phenomic prediction under drought in faba bean was strongly shaped by trait type, predictor composition, and temporal design. Combined VI + 3D predictors were most effective for integrative traits, whereas several component traits were predicted equally well or better by 3D information alone. These findings highlight the potential of scanner-based multisensor phenotyping to support drought-related selection in faba bean breeding.
Sylvain Poque · Ulrika Carlson-Nilsson · Muhammad Omer · Anna Palmé · Ingunn M. Vågen · G. Poulsen · Matti W. Leino · Kristiina Himanen · Hamid Khazaeı
Abstract Faba bean ( Vicia faba L.) has great potential to contribute to sustainable agriculture and protein security globally but is known to be very sensitive to drought stress. Uncovering drought-resilient germplasm is critical for developing resilient cultivars and advancing our understanding of the mechanisms underlying stress adaptation. However, high-throughput plant phenotyping under stress conditions remain a major bottleneck in crop genetics and breeding programs. In this study, a multi-sensor indoor phenotyping platform was used to assess 44 faba bean genotypes under water deficit conditions. Standardized, monitored stress conditions were achieved by watering-by-weighing for drought onset, duration, and intensities allowing genotype-level comparisons. The genotypes showed a range of stress responses in growth and physiology, including traits such as plant height, biomass, water use efficiency (WUE), and chlorophyll fluorescence parameters. Digital biomass, derived from combined top- and side-view plant imaging, was strongly correlated with biological biomass at the experimental endpoint, validating its use as a non-destructive proxy for growth assessment in faba bean. Time-resolved generalized additive modelling further revealed genotype-specific differences in the timing and magnitude of water deficit response. Genotypes that maintained growth and WUE under water deficit conditions may serve as valuable pre-breeding materials for development of drought-adapted faba bean.
Faba bean ( Vicia faba L.) is a key crop for sustainable agriculture in temperate cropping systems due to its nitrogen-fixing ability and high protein content, but its productivity is increasingly threatened by drought stress driven by climate change. Precise phenotyping under semicontrolled conditions is crucial for understanding drought responses. High-throughput precision phenotyping enables efficient evaluation of many genotypes, revealing detailed water-use patterns as a basis for breeding productive, drought-resilient cultivars. In this study, faba bean genotypes were grown in a precision phenotyping facility comprising 120-L containers filled with mineral soil to simulate field-like growth conditions. Each container was placed on a high-precision gravimetric scale to record water use in real time in relation to 3-D spectral image information. Precise measurement of genotype-specific transpiration behavior using gravimetric methods enabled detailed insights into the transpiration patterns of different genotypes in response to ambient temperature and humidity fluctuations throughout the day and night, and across the whole-life cycle. The results showed that total water use, water-use efficiency, and consequently yield were particularly influenced by specific transpiration parameters, such as the maximum transpiration rate and the vapor pressure deficit threshold at which stomatal conductance was declined. The results revealed genetically determined variation for transpiration responses to drought stress. Genotypes that reduced water loss earlier tended to achieve higher grain yields and use water more efficiently. The findings show that precise automated phenotyping can identify previously undiscovered genetic variation for breeding drought-tolerant faba bean varieties, which are crucial for ensuring productivity under increasingly water-limited conditions.
CONTEXT: Faba bean (Vicia faba L.) is a sustainable protein source, but in-season stresses such as heat, drought and diseases cause grain discolouration and shrivelling, leading to market downgrades. Grain quality assessments are only performed post-harvest, limiting growers’ ability to manage quality risks proactively on-farm. To address this limitation, this study explored the potential of in-season hyperspectral sensing as a non-destructive, data-driven tool for early grain quality assessment. AIMS: This study aimed to assess faba bean grain yield and quality pre-harvest by identifying optimal reproductive growth stage(s) and spectral regions linked to target grain traits. METHODS: Hyperspectral data were collected at five locations in Victoria, Australia across five critical reproductive growth stages: flowering (BBCH 65–69), podding (BBCH 70–79), pod fill (BBCH 80–82), pod maturity (BBCH 83–89), and crop senescence (BBCH 90–99). Partial least squares regression (PLSR) models were applied to canopy, leaf and pod level spectra to extract wavelength-trait relationships and identify predictive temporal windows for faba bean grain traits prediction prior to harvest. This approach enabled identification of both temporal (growth stage) and spectral (wavelength region) factors most informative for early trait prediction. Grain traits predicted include grain yield, harvest index, grain number, single grain weight, seed size index (SSI), grain protein content, seed coat brightness, redness and yellowness. KEY RESULTS: Canopy-level spectra provided the most reliable predictions. Harvest index (R² = 0.71, d-index = 0.75) and GPC (R² = 0.73, d-index = 0.76) were predicted as early as the flowering stage. The podding stage was optimal for predicting single grain weight (R² = 0.91, d-index = 0.76), SSI (R² = 0.71, d-index = 0.74), seed coat redness (R² = 0.68, d-index = 0.77) and yellowness (R² = 0.61, d-index = 0.68). Near-infrared (NIR) regions, 750–950 and 1000–1800 nm, were most informative for predicting grain quality traits. CONCLUSION: These findings demonstrate the potential of integrating hyperspectral sensing with chemometric modelling to enable pre-harvest prediction of faba bean grain agronomic and quality traits. IMPLICATIONS AND IMPACTS: Hyperspectral sensing as a precision agriculture application can mitigate on-farm grain quality downgrade risks by supporting early, data-driven harvest management decisions that maximise growers’ profitability and sustainability.
István Nagy · Peter Skov Kristensen · Elesandro Bornhofen · Marta Malinowska · Linda Kærgaard Nielsen · A. Schiemann · Niels Rolund · Stig Uggerhøj Andersen · Torben Asp
Abstract Protein-rich leguminous plants, such as faba bean and white clover are prospectively interesting crops in the North-European countries for reducing dependence on soybean import. Significant expansion of the production area of leguminous crops is challenged by the sub-optimal climatic conditions in this region, especially by the increasing probability of year-to-year fluctuation of extreme weather conditions due to global climate change. To overcome these challenges, development of new climate-resilient varieties suitable for growing under Northern-European conditions are needed. Root architecture and early root development, as well as the availability of efficient root phenotyping technologies are crucial factors of advancing in breeding of adequate varieties. We report a study of a simple and affordable screening technology of early root development using rhizoboxes in connection with semi-automated image analysis and provide a conceptual pipeline for estimation of Genomic Estimated Breeding Values (GEBVs) and correlating greenhouse and field phenotype data. Based on bivariate models, high genetic correlation (r=0.83) could be detected between total root length values recorded in greenhouse rhizobox experiments and field grain yield in faba bean. In white clover, moderately positive genetic correlation (r=0.17) between estimated breeding values of rhizobox-detected total root length and field yield could be identified. Our results suggest that phenotyping and selection of early root development components could potentially be useful in breeding programs to increase the genetic gain for field yield.
Faba bean ( Vicia faba L.) is a key protein crop, but its cultivation and yield stability are hindered by a number of environmental stresses. Stomata regulate gas exchange between the plant and atmosphere, playing a central role in photosynthesis and mediating plant responses to a wide range of environmental stressors. This study aimed to investigate variations in photosynthetic regulation in faba bean, and to examine leaf temperature and the response to short-term acute ozone (O₃) exposure as proxies for stomatal function. Here, we used a high-throughput plant phenotyping (HTPP) platform to screen 196 faba bean genotypes for photosynthetic and stomatal function under controlled conditions. A subset of extreme genotypes, identified based on relative leaf tempreture from the initial screening, was exposed to a 450 ppb O₃ treatment. Our results revealed strong positive relationship between photosynthetic efficiency and relative leaf temperature. A three-fold difference in relative leaf temperature was observed among genotypes. The O₃ treatment caused signicantly less damage in genotypes with higher leaf temperature compared to those with lower leaf temperature (p < 0.001). By combining a HTPP platform with elevated O₃ stress treatment, we identified faba bean genotypes with contrasting stomatal responses to the O₃ exposure. Our results advance understanding of the regulation mechanisms of photosynthetic light reactions and the role of stomatal function in modulating faba bean responses to environmental stressors. • High-throughput phenotyping reveals large variation in leaf temperature among faba bean genotypes. • Leaf temperature strongly affects photosynthetic regulation in faba bean. • The tested genotypes with higher leaf temperatures display increased ozone tolerance.
Faba bean (Vicia faba L.) is a valuable legume crop with high protein content and the ability to fix atmospheric nitrogen through symbiotic bacteria in its root nodules, contributing significantly to both human nutrition and agricultural sustainability. Chlorophyll concentration in leaves serves as a reliable indicator of nitrogen status and photosynthetic capacity, while biomass production reflects overall plant growth and resource use efficiency. This study aims to develop a nondestructive and accurate method for simultaneously estimating chlorophyll content meter (SPAD) and above-ground biomass in faba bean using three-dimensional (3D) photogrammetric imaging combined with deep learning techniques. Point cloud data were obtained from hand-held camera scans of five faba bean genotypes and processed using the PointNet neural network architecture. Results showed that SPAD estimation achieved high accuracy (7.52% relative error) based solely on 3D structural features, while biomass prediction benefited from the integration of real and synthetic datasets, reducing relative error significantly from 24.15 % to 18.04 %. The study highlights the potential of 3D imaging and point cloud-based photogrammetry as effective tools for plant phenotyping, offering a scalable and non-invasive approach for monitoring physiological traits and genotype performance in faba bean.
Crop monitoring is paramount to ensure effective and sustainable agricultural practices. These activities provide crucial information about crop health, development, and yield, enabling farmers to make informed decisions and enhance their farming practices. However, deep learning has proven to be a vital tool. It allows the automated analysis of vast agricultural data, delivering precise and timely information for proactive crop management and resource allocation decision-making. Based on an enhanced convolutional neural network model, the proposed framework focuses on detecting three key growth stages in Vicia faba L. cultivation within challenging and intricate environments. The dataset utilized in this study comprises images representing diverse developmental phases of crops collected through Unmanned Aerial Vehicles (UAVs) at an agricultural farm during different periods. Four distinct models within the framework were evaluated based on classification accuracy, mean average precision (mAP), and F1 score. The results indicate that the model with the highest classification accuracy reached 91.6\%, with a commendable mAP of 90.7\%. In contrast, the model with the lowest accuracy achieved a precision of 88.2\%. The empirical validation of the framework in a complex agricultural environment aligns seamlessly with the demands of modern farming operations, demonstrating notable improvements in precision and reliability.
Aphids hide under leaves, reproduce rapidly, and require early detection to prevent crop damage, disease transmission, and ensure effective pest management. This study presents a novel approach for aphid detection by utilizing hyperspectral imaging, multivariate classification methods and spectral information divergence (SID) analyses. The hyperspectral images average spectrum ( n = 336) showed significant differences between healthy and infested leaves. Time-series classification was performed over 14 days after infestation using four distinct machine learning algorithms. Early-stage infection detection may not relate to internal physiological alterations within the leaf but rather to the physical presence of the aphid behind the leaf, obstructing subtle physiological signatures. Implementation of spectral endmembers in the VIS-NIR reference spectrum led to the identification of an informative abundance SID map within the 710-825 nm range, useful for further classification. Machine learning classification resulted in support vector machines achieving 99.20 accuracy. Using random forest, twenty-two most important variables found effective in boosting classifier performance. The selected model also extended to real-world scenarios by testing progressing infestation patterns over 14 days on independent data sets, confirming the system's reliability. Signal normal variant pre-treatment with partial least squares regression was effective in the estimation of aphid populations, achieving a 0.81 coefficient of determination ( R 2 ) and a 10.29 root-mean-square error of prediction for test datasets. In conclusion, the proposed method was able to successfully detect aphid colony infestation, both earlier and in locations that are invisible during standard human inspection.
Abstract Faba bean ( Vicia faba L.) has great potential to contribute to sustainable agriculture and protein security globally. However, it is known to be very sensitive to droughts, which can severely impact yield. Uncovering drought-resilient germplasm is critical for developing resilient cultivars and advancing our understanding of the mechanisms underlying stress adaptation. However, reliable phenotyping of water stress responses remains a significant bottleneck in crop genetics and breeding programs. Overcoming this bottleneck requires high-throughput phenotyping platforms. In this study, we used an indoor image-based phenotyping facility, the National Plant Phenotyping Infrastructure at the University of Helsinki. The facility incorporates cutting-edge imaging technologies such as top- and side-view digital imaging for assessment of growth and development, as well as chlorophyll fluorometry for the detection of physiological responses. In this study, 44 faba bean accessions were subjected to early-stage water stress via weight-based water-holding capacity. The accessions presented a range of responses to water stress across the studied traits, including plant height, total canopy area, digital biomass, and water use efficiency. Our results also revealed a strong correlation between digital biomass and biological biomass. Here, we demonstrate the potential of a fully automated indoor phenotyping facility for screening a relatively large faba bean germplasm collection under well watered and water stressed conditions. Accessions that maintained growth and physiological performance under water stress conditions in this study may serve as valuable pre-breeding materials for the development of drought-adapted faba beans.
Abstract Background Faba bean ( Vicia faba L.) is a key crop for sustainable agriculture in temperate cropping systems due to its nitrogen-fixing ability and high protein content, but its productivity is increasingly threatened by drought stress driven by climate change. Precise phenotyping under semi-controlled conditions is crucial for understanding drought responses. High-throughput precision phenotyping enables efficient evaluation of many genotypes, revealing detailed water-use patterns as a basis for breeding productive, drought-resilient cultivars. In this study, faba bean genotypes were grown in a precision phenotyping facility comprising 120-liter containers filled with mineral soil to simulate field like growth conditions. Each container was placed on a high-precision gravimetric scale to record water use in real-time in relation to 3-dimensional spectral image information. Results Precise measurement of genotype-specific transpiration behavior using gravimetric methods enabled detailed insights into the transpiration patterns of different genotypes in response to ambient temperature and humidity fluctuations throughout the day, night and across the whole life cycle. The results showed that total water use, water use efficiency, and consequently yield were particularly influenced by specific transpiration parameters, such as the maximum transpiration rate and the vapor pressure deficit threshold at which stomatal conductance was interrupted. Conclusion The results revealed genetically determined variation for transpiration responses to drought stress. Genotypes that reduced water loss earlier tended to achieve higher grain yields and use water more efficiently. The findings show that precise automated phenotyping can identify previously undiscovered genetic variation for breeding of drought-tolerant faba bean varieties, which are crucial for ensuring productivity under increasingly water-limited conditions.
Stomata are vital for CO 2 and water vapor exchange, with guard cells' aperture and ultrastructure highly responsive to environmental cues. However, traditional methods for studying guard cell ultrastructure, which rely on chemical fixation and embedding, often distort cell morphology and compromise membrane integrity. In contrast, plunge-freezing in liquid ethane rapidly preserves cells in a near-native vitreous state for cryogenic electron microscopy. Using this approach, we applied Cryo-Focused Ion Beam-Scanning Electron Microscopy (cryo-FIB-SEM) to study the guard cell ultrastructure of Vicia faba, a higher plant model chosen for its sensitivity to external factors and ease of epidermis isolation, advancing beyond previous cryo-FIB-SEM applications in lower plant algae. The results firstly introduced cryo-FIB-SEM volume imaging, enabling subcellular ultrastructure visualization of higher plants like V. faba in a vitrified, unaltered state. 3D models of organelles such as stromules, chloroplast protrusions, chloroplasts, starch granules, mitochondria, and vacuoles were reconstructed from cryo-FIB-SEM volumetric data, with their surface area and volume initially determined using manual segmentation. Future studies using this near-native volume imaging technique hold promise for investigating how environmental factors like drought or salinity influence stomatal behavior and the morphology of guard cells and their organelles.
Abstract Stomata regulate gas exchange between plants and the atmosphere, but analysing their morphology is challenging due to anatomical variability and artefacts during image acquisition. Deep learning (DL) can address these challenges but often requires large and diverse datasets, which are costly and error prone to produce. Generative adversarial networks (GANs) offer a solution by generating artificial data via unsupervised learning. However, GANs often suffer from problems including mode collapse, vanishing gradients, and network failure, particularly with small datasets. Here, we present StomaGAN, a deep convolutional GAN (DCGAN) with tailored modifications to address common GAN issues. We collected 559 stomatal impressions of field, or faba bean (Vicia faba) consisting of ~3000 stoma, 80% of which were used to train StomaGAN. Evaluation metrics, including generator and discriminator loss progression and a mean Fréchet Inception Distance (FID) score of 61.4 across eight experimental runs confirm successful training. To validate StomaGAN, we generated artificial images to train a deep convolutional neural network (DCNN) based on the DeepLabV3 framework for stomata detection from real, unseen images. The DCNN achieved a mean Interception over Union (IoU) of 0.95 on artificial training images and 0.91 on real, unseen, images across varying magnifications. Our results demonstrate that StomaGAN effectively generates high-quality synthetic datasets, enabling reliable stomatal detection and enhancing phenotypic analysis. This approach reduces the need for extensive manual data collection and simplifies complex morphological assessments.
Efficient crop disease management shows great promise in optimizing the agricultural industry. Accurate identification of infection levels is crucial for implementing effective and efficient disease treatments. However, accurately identifying and locating crop diseases in complex, unstructured field environments remain challenging. This necessitates the utilization of large volumes of annotated data. This research paper comprehensively evaluates deep transfer learning techniques for identifying the degree of rust disease infection in Vicia faba L. production systems. We curate a vast dataset comprising images captured under natural lighting conditions and at different growth stages of the crop under study. We propose a deep learning model based on Neural Architecture Search (NAS) specifically designed for early detection and accurate classification of disease levels in crops. We compare the performance of our proposed model with nine other deep learning models using transfer learning. Remarkably, transfer learning based on the NAS method achieves high classification accuracy, consistently exceeding 90.84% F1 scores. Moreover, all models exhibit short training times, requiring less than 3 hours. Among the evaluated models, our NAS-based model emerges as the top performer, highlighting the importance and effectiveness of this method in developing state-of-the-art models. It achieves a mean average precision of 94.10% and an impressive overall recall of 96.96%. These results significantly contribute to developing robust and accurate disease management strategies, paving the way for improved agricultural practices and increased crop yields. Our approach enables early disease detection and precise classification, leveraging deep transfer learning and facilitating timely interventions and optimized treatments. With the help of this study, we can now better utilize cuttingedge agricultural technology, paving the way for sustainable crop production in the future.
Chocolate spot (CS), caused by Botrytis fabae, is one of the most destructive fungal diseases affecting faba bean (Vicia faba L.) globally. This study evaluated 33 faba bean cultivars across two locations and over 2 years to assess genetic resistance and the effect of fungicide application on CS progression. The utility of unmanned aerial vehicle–mounted multispectral camera for disease monitoring was examined. Significant variability was observed in cultivar susceptibility, with Bolivia exhibiting the highest level of resistance and Louhi, Sampo, Vire, Merlin, Mistral, and GL Sunrise proving highly susceptible. Fungicide application significantly reduced CS severity and improved yield. Analysis of canopy spectral signatures revealed the near‐infrared and red edge bands, along with enhanced vegetation index (EVI) and soil adjusted vegetation index, as most sensitive to CS infection, and they had a strong negative correlation with CS severity ranging from −0.51 to −0.71. In addition, EVI enabled early disease detection in the field. Support vector machine accurately classified CS severity into four classes (resistant, moderately resistant, moderately susceptible, and susceptible) based on spectral data with higher accuracy after the onset of disease compared to later in the season (accuracy 0.75–0.90). This research underscores the value of integrating resistant germplasm, sound agronomic practices, and spectral monitoring for effectively identification and managing CS disease in faba bean.
Faba bean is a global food legume crop, and it is essential to accurately and timely determine its plant height, above-ground biomass (fresh and dry weight) and yield for enhancing cultivation practices and planning the next planting season. Traditional ground sampling is a time-consuming and labor-intensive approach. However, the utilization of an unmanned aerial vehicle (UAV) as a high-throughput technique offers a promising alternative strategy for estimating crop phenotypic traits. In this study, a two-year experiment was conducted from 2020 to 2022, where UAV-based multimodal data were collected using red-green-blue, multispectral and thermal infrared sensors. The variables derived from these three sensors and their combinations were used to estimate the fresh weight, dry weight and yield of faba bean based on extreme gradient boosting (XGBoost), random forest, multiple linear regression and k-nearest neighbor algorithms. The following findings were obtained: (1) The use of the maximum percentile crop surface model resulted in the highest estimation accuracy for faba bean plant height. (2) Fusion data from multiple sensors increased the estimation accuracy of faba bean fresh weight, dry weight and yield, the coefficient of determination (R²) improved by 14.22%, 1.45%, and 18.76%, respectively, compared with the best estimation accuracy of a single sensor. (3) The XGBoost algorithm outperformed the other algorithms in estimating fresh weight, dry weight and yield of faba bean. These results demonstrate that multiple sensors and appropriate algorithms can be used to effectively estimate faba bean phenotypic traits and provide valuable insights for agricultural remote sensing research.
Stomata are vital for CO2 and water vapor exchange, with guard cells’ aperture and ultrastructure highly responsive to environmental cues. However, traditional methods for studying guard cell ultrastructure, which rely on chemical fixation and embedding, often distort cell morphology and compromise membrane integrity, leaving no suitable methodology until now. In contrast, plunge-freezing in liquid ethane rapidly preserves cells in a near-native vitreous state for cryogenic electron microscopy. Using this approach, we applied Cryo-Focused Ion Beam-Scanning Electron Microscopy (cryo- FIB-SEM) to study the guard cell ultrastructure of Vicia faba , a higher plant model chosen for its sensitivity to external factors and ease of epidermis isolation, advancing beyond previous cryo-FIB-SEM applications in lower plant algae. The results firstly introduced cryo-FIB-SEM volume imaging, enabling subcellular ultrastructure visualization of higher plants like V. faba in a vitrified, unaltered state. 3D models of organelles such as stromules, chloroplast protrusions, chloroplasts, starch granules, mitochondria, and vacuoles were reconstructed from cryo-FIB-SEM volumetric data, with their surface area and volume initially determined using manual segmentation. Future studies using this near-native volume imaging technique hold promise for investigating how environmental factors like drought or salinity influence stomatal behavior and the morphology of guard cells and their organelles.
Accurately and economically estimated crop above-ground biomass (AGB) and bean yield (BY) are critical for cultivation management in precision agriculture. Unmanned aerial vehicle (UAV) platforms have shown great potential in crop AGB and BY estimation due to their ability to rapidly acquire remote sensing data with high temporal–spatial resolution. In this study, a low-cost and consumer-grade camera mounted on a UAV was adopted to acquire red–green–blue (RGB) images, which were then combined with ensemble learning to estimate faba bean AGB and BY. The following results were obtained: (1) The faba bean plant height derived from UAV RGB images presented a strong correlation with the ground measurement (R² = 0.84, RMSE = 63.6 mm). (2) The accuracy of BY estimation (R² = 0.784, RMSE = 0.460 t ha⁻¹, NRMSE = 14.973%) based on RGB images was higher than the accuracy of AGB estimation (R² = 0.618, RMSE = 0.606 t ha⁻¹, NRMSE = 16.746%). (3) The combination of three variables (vegetation index, structural information, textural information) improved the AGB and BY estimation accuracy. (4) The AGB and BY estimation performance were best for the mid bean-filling stage. (5) The ensemble learning model provided higher AGB and BY estimation accuracy than the five base learners (k-nearest neighbor, support vector machine, ridge regression, random forest and elastic net models). These results indicate that UAV RGB images combined with machine learning algorithms, particularly ensemble learning models, can provide relatively accurate faba bean AGB (R² = 0.683, RMSE = 0.568 t ha⁻¹, NRMSE = 15.684%) and BY (R² = 0.854, RMSE = 0.390 t ha⁻¹, NRMSE = 12.693%) estimation and considerably contribute to the high-throughput phenotyping study of food legumes.
Near-infrared spectroscopy (NIRS) provides a high-throughput phenotyping technique to assist breeding for improved faba bean seed quality. We combined chemical analysis of protein, oil content (and composition) with NIRS through chemometrics, employing Partial Least Squares (PLS), Elastic Net (EN), Memory-based Learning (MBL), and Bayes B (BB) as prediction models. Protein was the most reliably predicted trait (R2 = 0.96–0.98) across field trials, followed by oil (R2 = 0.82–0.86) and oleic acid (R2 = 0.31–0.68). Samples for training the models were selected using K-means clustering. The optimal statistical approach for prediction was compound-specific: PLS for protein (Root Mean Squared Error - RMSE = 0.46), BB for oil (RMSE = 0.067), and EN for oleic acid content (RMSE = 2.83). Reduced training set simulations revealed different effects on prediction accuracy depending on the model and compound. Several NIR regions were pinpointed as highly informative for the compounds, using the shrinkage and variable selection capabilities of EN and BB.
BACKGROUND: Image-based crop growth modeling can substantially contribute to precision agriculture by revealing spatial crop development over time, which allows an early and location-specific estimation of relevant future plant traits, such as leaf area or biomass. A prerequisite for realistic and sharp crop image generation is the integration of multiple growth-influencing conditions in a model, such as an image of an initial growth stage, the associated growth time, and further information about the field treatment. While image-based models provide more flexibility for crop growth modeling than process-based models, there is still a significant research gap in the comprehensive integration of various growth-influencing conditions. Further exploration and investigation are needed to address this gap. METHODS: We present a two-stage framework consisting first of an image generation model and second of a growth estimation model, independently trained. The image generation model is a conditional Wasserstein generative adversarial network (CWGAN). In the generator of this model, conditional batch normalization (CBN) is used to integrate conditions of different types along with the input image. This allows the model to generate time-varying artificial images dependent on multiple influencing factors. These images are used by the second part of the framework for plant phenotyping by deriving plant-specific traits and comparing them with those of non-artificial (real) reference images. In addition, image quality is evaluated using multi-scale structural similarity (MS-SSIM), learned perceptual image patch similarity (LPIPS), and Fréchet inception distance (FID). During inference, the framework allows image generation for any combination of conditions used in training; we call this generation data-driven crop growth simulation. RESULTS: Experiments are performed on three datasets of different complexity. These datasets include the laboratory plant Arabidopsis thaliana (Arabidopsis) and crops grown under real field conditions, namely cauliflower (GrowliFlower) and crop mixtures consisting of faba bean and spring wheat (MixedCrop). In all cases, the framework allows realistic, sharp image generations with a slight loss of quality from short-term to long-term predictions. For MixedCrop grown under varying treatments (different cultivars, sowing densities), the results show that adding these treatment information increases the generation quality and phenotyping accuracy measured by the estimated biomass. Simulations of varying growth-influencing conditions performed with the trained framework provide valuable insights into how such factors relate to crop appearances, which is particularly useful in complex, less explored crop mixture systems. Further results show that adding process-based simulated biomass as a condition increases the accuracy of the derived phenotypic traits from the predicted images. This demonstrates the potential of our framework to serve as an interface between a data-driven and a process-based crop growth model. CONCLUSION: The realistic generation and simulation of future plant appearances is adequately feasible by multi-conditional CWGAN. The presented framework complements process-based models and overcomes their limitations, such as the reliance on assumptions and the low exact field-localization specificity, by realistic visualizations of the spatial crop development that directly lead to a high explainability of the model predictions.
Reproduction assets foundThe paper's authors explicitly state that source code and links to the phenotyping datasets (Arabidopsis, GrowliFlower, MixedCrop) are publicly available in their GitHub repository, which implements the multi-conditional CWGAN crop growth simulation and growth estimation framework.Code · publicSource code and links to the datasets are publicly available at https://github.com/luked12/crop-growth-cgan .Open asset ↗luked12/crop-growth-cganlines:216-253Plant phenotyping relevance matchCrossref · checked 14 Sept 2026
Background: Legumes, such as lentils, field peas, Faba beans and chickpeas, are high in vitamins, fiber, important minerals and protein and can help avoid obesity and cardiovascular illnesses. They also contribute to ecosystem services, such as nitrogen fixation and resilience to environmental stresses. Despite a 60% increase in global pulse production from 2000 to 2021, a demand-supply gap, especially in South Asia, raises concerns about nutritional access. Since illnesses are currently an issue to the food security of faba beans, machine learning is required for efficient disease identification. Methods: This research employs Convolutional Neural Networks (CNNs) for robust Faba bean leaf disease identification. The CNN model is trained with diverse images representing specific diseases. The study focuses on diseases like Chocolate Spot, Faba Bean Gall, Rust and Healthy leaves. Image processing involves resizing, grayscale conversion and labeling. The CNN architecture includes eight convolutional layers, four max-pooling layers and three dropout layers. The model is trained using 80% of the dataset, validated using 20% and tested for accuracy. Result: The CNN model achieves an accuracy of 99.37% during training and 89.69% during validation after 75 epochs. Confusion matrix and classification report illustrate the model’s performance. It shows high precision, recall and F1 scores for each class, indicating balanced performance. Chocolate Spot and Rust exhibit the highest precision and F1 scores. The overall accuracy is 91%, comparable to other studies on Faba bean disease detection. The study presents a CNN-based disease identification system for Faba beans, demonstrating high accuracy and balanced performance across different diseases. The model’s effectiveness is comparable to other advanced techniques. The research highlights the potential of machine learning in optimizing disease management for Faba beans. Future work could explore a broader range of diseases and incorporate hybrid machine learning algorithms for further improvement.
Background The detection of internal defects in seeds via non-destructive imaging techniques is a topic of high interest to optimize the quality of seed lots. In this context, X-ray imaging is especially suited. Recent studies have shown the feasibility of defect detection via deep learning models in 3D tomography images. We demonstrate the possibility of performing such deep learning-based analysis on 2D X-ray radiography for a faster yet robust method via the X-Robustifier pipeline proposed in this article. Results 2D X-ray images of both defective and defect-free seeds were acquired. A deep learning model based on state-of-the-art object detection neural networks is proposed. Specific data augmentation techniques are introduced to compensate for the low ratio of defects and increase the robustness to variation of the physical parameters of the X-ray imaging systems. The seed defects were accurately detected (F1-score >90%), surpassing human performance in computation time and error rates. The robustness of these models against the principal distortions commonly found in actual agro-industrial conditions is demonstrated, in particular, the robustness to physical noise, dimensionality reduction and the presence of seed coating. Conclusion This work provides a full pipeline to automatically detect common defects in seeds via 2D X-ray imaging. The method is illustrated on sugar beet and faba bean and could be efficiently extended to other species via the proposed generic X-ray data processing approach (X-Robustifier). Beyond a simple proof of feasibility, this constitutes important results toward the effective use in the routine of deep learning-based automatic detection of seed defects.
Measuring canopy height is important for phenotyping as it has been identified as the most relevant parameter for the fast determination of plant mass and carbon stock, as well as crop responses and their spatial variability. In this work, we develop a low-cost tool for measuring plant height proximally based on an ultrasound sensor for flexible use in static or on-the-go mode. The tool was lab-tested and field-tested on crop systems of different geometry and spacings: in a static setting on faba bean (Vicia faba L.) and in an on-the-go setting on chia (Salvia hispanica L.), alfalfa (Medicago sativa L.), and wheat (Triticum durum Desf.). Cross-correlation (CC) or a dynamic time-warping algorithm (DTW) was used to analyze and correct shifts between manual and sensor data in chia. Sensor data were able to reproduce with minor shifts in canopy profile and plant status indicators in the field when plant heights varied gradually in narrow-spaced chia (R2 = 0.98), faba bean (R2 = 0.96), and wheat (R2 = up to 0.99). Abrupt height changes resulted in systematic errors in height estimation, and short-scale variations were not well reproduced (e.g., R2 in widely spaced chia was 0.57 to 0.66 after shifting based on CC or DTW, respectively)). In alfalfa, ultrasound data were a better predictor than NDVI (Normalized Difference Vegetation Index) for Leaf Area Index and biomass (R2 from 0.81 to 0.84). Maps of ultrasound-determined height showed that clusters were useful for spatial management. The good performance of the tool both in a static setting and in the on-the-go setting provides flexibility for the determination of plant height and spatial variation of plant responses in different conditions from natural to managed systems.
Abstract Unmanned aerial vehicles (UAVs) equipped with high-resolution imaging sensors have shown great potential for plant phenotyping in agricultural research. This study aimed to explore the potential of UAV-derived red–green–blue (RGB) and multispectral imaging data for estimating classical phenotyping measures such as plant height and predicting yield and chlorophyll content (indicated by SPAD values) in a field trial of 38 faba bean (Vicia faba L.) cultivars grown at four replicates in south-eastern Norway. To predict yield and SPAD values, Support Vector Regression (SVR) and Random Forest (RF) models were utilized. Two feature selection methods, namely the Pearson correlation coefficient (PCC) and sequential forward feature selection (SFS), were applied to identify the most relevant features for prediction. The models incorporated various combinations of multispectral bands, indices, and UAV-based plant height values at four different faba bean development stages. The correlation between manual and UAV-based plant height measurements revealed a strong agreement with a correlation coefficient (R2) of 0.97. The best prediction of SPAD value was achieved at BBCH 50 (flower bud present) with an R2 of 0.38 and RMSE of 1.14. For yield prediction, BBCH 60 (first flower open) was identified as the optimal stage, using spectral indices yielding an R2 of 0.83 and RMSE of 0.53 tons/ha. This development stage presents an opportunity to implement targeted management practices to enhance yield. The integration of UAVs equipped with RGB and multispectral cameras, along with machine learning algorithms, proved to be an accurate approach for estimating agronomically important traits in faba bean. This methodology offers a practical solution for rapid and efficient high-throughput phenotyping in faba bean breeding programs.
Early and high-throughput estimations of the crop harvest index (HI) are essential for crop breeding and field management in precision agriculture; however, traditional methods for measuring HI are time-consuming and labor-intensive. The development of unmanned aerial vehicles (UAVs) with onboard sensors offers an alternative strategy for crop HI research. In this study, we explored the potential of using low-cost, UAV-based multimodal data for HI estimation using red-green-blue (RGB), multispectral (MS), and thermal infrared (TIR) sensors at 4 growth stages to estimate faba bean (Vicia faba L.) and pea (Pisum sativum L.) HI values within the framework of ensemble learning. The average estimates of RGB (faba bean: coefficient of determination [R2] = 0.49, normalized root-mean-square error [NRMSE] = 15.78%; pea: R2 = 0.46, NRMSE = 20.08%) and MS (faba bean: R2 = 0.50, NRMSE = 15.16%; pea: R2 = 0.46, NRMSE = 19.43%) were superior to those of TIR (faba bean: R2 = 0.37, NRMSE = 16.47%; pea: R2 = 0.38, NRMSE = 19.71%), and the fusion of multisensor data exhibited a higher estimation accuracy than those obtained using each sensor individually. Ensemble Bayesian model averaging provided the most accurate estimations (faba bean: R2 = 0.64, NRMSE = 13.76%; pea: R2 = 0.74, NRMSE = 15.20%) for whole growth stage, and the estimation accuracy improved with advancing growth stage. These results indicate that the combination of low-cost, UAV-based multimodal data and machine learning algorithms can be used to estimate crop HI reliably, therefore highlighting a promising strategy and providing valuable insights for high spatial precision in agriculture, which can help breeders make early and efficient decisions.
Abstract Faba bean is a vital legume crop, and its early yield estimation can improve field management practices. In this study, unmanned aerial system (UAS) hyperspectral imagery was used for the first time to estimate faba bean yield early. Different basic algorithms, including random forest (RF), support vector machine (SVM), k-nearest neighbor (KNN), partial least squares regression (PLS), and eXtreme Gradient Boosting (XGB), were employed along with stacking ensemble learning to construct the faba bean yield model and investigate factors influencing model accuracy. The results are as follows: when using the same algorithm and growth period, integrating texture information into the model improved the estimation accuracy compared to using spectral information alone. Among the base models, the XGB model performed the best in the context of growth period consistency. Moreover, the stacking ensemble significantly improved model accuracy, yielding satisfactory results, with the highest model accuracy (R 2 ) reaching 0.76. Model accuracy varied significantly for models based on different growth periods using the same algorithm. The accuracy of the model gradually improved during a single growth period, but the rate of improvement decreased over time. Data fusion of growth period data helped enhance model accuracy in most cases. In conclusion, combining UAS-based hyperspectral data with ensemble learning for early yield estimation of faba beans is feasible, therefore, this study would offer a novel approach to predict faba bean yield.
Traditional field-based techniques for phenotyping of crops are based on visual assessment which are subjective and time consuming. A high throughput automated technique using an unmanned aerial vehicle (UAV) with a multispectral image (MSI) camera was used to investigate the correlation between markers of winter bean crop development with eventual crop yield. A simplified approach has been developed using different vegetation indices i.e. NDVI, GNDVI and NDRE, coupled with an iso-cluster classification method to monitor plant characteristics across all growing stages. The UAV-MSI data could then be incorporated into a yield estimator model to estimate the winter bean seed yield. The NDVI approach showed the greatest correlation between the modelled seed yield and the actual seed yield determined on two separate occasions (R2 = 0.84 and R2 = 0.87). In addition, GNDVI and NDRE were a better estimator of seed yield for areas with dense vegetation. These are hence shown to be able to monitor a winter bean harvest in an efficient and timely manner.
To overcome threats to agro-ecosystems, such as a dramatic species decline, an ecological intensification in crop production is needed. One possible strategy is the simultaneous cultivation of legume and cereal plants in a mixed arrangement, namely mixed cropping. Cereal-legume crop mixtures may benefit from diversity effects, i.e. improved use of environmental resources such as light, water and nitrogen. Thus, mixtures have shown to result in higher land productivity with respect to grain yield compared to sole cropping. However, mixture systems are complex and difficult to study due to dynamic species interactions and their heterogeneous canopy structures. To better understand structural and functional diversity effects in a mixed cropping system, we non-invasively studied two crops in a field trial in 2021 and 2022. Here, different genotypes of faba bean (Vicia faba L.) and spring wheat (Triticum aestivum L.) were combined in six legume-cereal mixtures. The 1:1 mixtures were compared to each other and against the respective sole crops. To study structural and functional diversity effects in mixtures, we applied proximal and remote sensing tools. We characterized photosynthesis-related plant traits derived from hyperspectral and solar-induced fluorescence (SIF) data recorded with ground-based and airborne sensors. The high-performance airborne spectrometer HyPlant was used to acquire SIF image data with 1 m spatial resolution. Additionally, we collected hyperspectral and SIF point measurements with the mobile field sensor system FloX on different dates during the two growing seasons. We found that HyPlant and FloX datasets of different mixtures and crop types collected in mid-June showed significantly different levels of far-red SIF emission efficiency (εF) (p
Abstract The stomata on the leaf surface are mainly responsible for the material exchange between the internal and external environments of the plant, a large number of methods have been proposed to automatically measure the distribution position and number of stomatal, but few methods could achieve both stomatal count and open/closed‐state judgment. Therefore, this study proposes an automatic detection method for leaf stomatal morphology analysis based on an attention mechanism and deep learning. In order to obtain more stomatal feature information and send it to the network for learning, the proposed method adds a coordinate attention (CA) mechanism to the YOLOV5 backbone part. At the same time, in order to avoid the overfitting of the model during the training process, the authors added the training trick of label smoothing. Finally, the detection ability of the proposed method for stomata is verified on the broad bean leaves stomata dataset. The experimental results show that our method achieves a detection accuracy of 0.934 and an mAP of 0.968. By comparing with other state‐of‐the‐art algorithms, the detection capability of our method has been significantly improved. The generalization of the model is verified on the wheat leaf stomatal dataset. The experimental results show that our method can achieve a detection accuracy of 0.894 and an mAP of 0.907.
Reproduction assets foundThe paper's broad bean/wheat leaf stomata microscopy image dataset (951 broad bean + 160 wheat images with YOLO-format annotations) is openly deposited on Zenodo per the data availability statement. No code or trained model deposit is explicitly stated.Dataset · publicOF INTEREST
problems, and its indicators are better than the six comparison The authors declare that there are no conflict of interests, we do
algorithms above. not have any possible conflicts of interest.
DATA AVAILABILITY STATEMENT
5 CONCLUSIONS The data that support the findings of this study are openly
available in zendo at https://doi.org/10.5281/zenodo.6302925.
In order to better detect and count the position, number, and
open/closed-status of stomata in plant leaves, we introduce a AUTHOR CONTRIBUTIONS
modified end-to-end target detection model YOLOv5 in this Xin Li: Conceptualization; Data curation; Formal analysis;
study. In order to improve the ability of YOLOv5s model to InvestiOpen asset ↗zenodo · 10.5281/zenodo.6302925pdf-layout-page:9 lines:1-55Plant phenotyping relevance matchCrossref · checked 8 Sept 2026
Abstract Background Faba bean is an important legume crop in the world. Plant height and yield are important traits for crop improvement. The traditional plant height and yield measurement are labor intensive and time consuming. Therefore, it is essential to estimate these two parameters rapidly and efficiently. The purpose of this study was to provide an alternative way to accurately identify and evaluate faba bean germplasm and breeding materials. Results The results showed that 80% of the maximum plant height extracted from two-dimensional red–green–blue (2D-RGB) images had the best fitting degree with the ground measured values, with the coefficient of determination (R 2 ), root-mean-square error (RMSE), and normalized root-mean-square error (NRMSE) were 0.9915, 1.4411 cm and 5.02%, respectively. In terms of yield estimation, support vector machines (SVM) showed the best performance (R 2 = 0.7238, RMSE = 823.54 kg ha −1 , NRMSE = 18.38%), followed by random forests (RF) and decision trees (DT). Conclusion The results of this study indicated that it is feasible to monitor the plant height of faba bean during the whole growth period based on UAV imagery. Furthermore, the machine learning algorithms can estimate the yield of faba bean reasonably with the multiple time points data of plant height.
In view of climate change, increasing soil salinity is expected worldwide. It is therefore important to improve prediction ability of plant salinity effects. For this purpose, brackish/saline irrigation water from two areas in central and coastal Tunisia was sampled. The water samples were classified as C3 (EC: 2.01-2.24 dS m -1 ) and C4 (EC: 3.46-7.00 dS m -1 ), indicating that the water was questionable and not suitable for irrigation, respectively. The water samples were tested for their genotoxic potential and growth effects on Vicia faba seedlings. Results showed a decrease in mitotic index (MI) and, consequently, growth parameters concomitant to the appearance of micronucleus (MCN) and chromosome aberrations when the water salinity increased. Salt ion concentration had striking influence on genome stability and growth parameters. Pearson correlation underlined the negative connection between most ions in the water inappropriate for irrigation (C4) and MI as well as growth parameters. MI was strongly influenced by Mg 2+ , Na + , Cl - , and to a less degree Ca 2+ , K + , and SO 4 2- . Growth parameters were moderately to weakly affected by K + and Ca 2+ , respectively. Re-garding MCN, a very strong positive correlation was found for MCN and K + . Despite its short-term application, the Vicia -MCN Test showed a real ability to predict toxicity induced by salt ions confirming that is has a relevant role in hazard identification and risk assessment of salinity effects.
Extreme temperatures at critical developmental phases reduce grain yield. Combinations of sowing date and cultivar that favour faster development reduce the likelihood of heat stress but increase the risk of frost at critical phases. Current models are unable to predict pulse yield in response to frost and heat, hence our focus on phenology. Our aim was to model phenological variation with sowing date and cultivar for lentil and faba bean against the climatic patterns of frost and heat in 45 Australian locations that spanned 29 °S-41 °S, 11−340 m.a.s.l., and 1−423 km to the coast.For both crops, modelled mean and standard deviation of time to flowering were close to actuals and mean prediction error was below 5%. Comparison of actual and modelled time to flowering returned: r = 0.89 (n = 121, P 34 °C) probabilities between 1957 and 2018 were used to estimate the date of 10 % frost probability and the date of 30 % heat probability as the boundaries of a frost-heat risk window for the critical period. Out of the 45 locations, 12 were frost-free but with risk of heat, 7 were heat-free but with risk of frost, 3 were frost- and heat-free, and 23 featured a window defined by both frost and heat boundaries. Frost variables discriminated locations more strongly than heat variables. Geographical patterns in thermal regimes emerged that were associated with latitude, altitude and continentality.Realised warming between 1957 and 2018 advanced the time to 200 °Cd after flowering and shortened the critical period in most locations, particularly in early-sown crops. Comparisons of the probability curves of frost and heat between 1957–1985 and 1986–2018 showed, with few exceptions, an asymmetry between delayed late frost (up to 44 d) and earlier heat onset (up to 11 d), with a narrowing of the frost-heat risk window from 46 to 90 d for the period 1957–1985 to 34–64 d for 1986–2018.We identified a dominant role of frost as (i) the main discriminating factor among geographically distinct locations, (ii) the main source of variation of the frost-heat window, and (iii) a putatively increased risk factor with climate change. Adaptation to frost in the critical period for yield is important for pulses despite warming trends. Increased frost tolerance can directly improve yield and indirectly contribute to reduce risk of heat and drought later in the season.
The traditional visual rating system is labor-intensive, time-consuming, and prone to human error. Unmanned aerial vehicle (UAV) imagery-based vegetation indices (VI) have potential applications in high-throughput plant phenotyping. The study objective is to determine if UAV imagery provides accurate and consistent estimations of crop injury from herbicide application and its potential as an alternative to visual ratings. The study was conducted at the Kernen Crop Research Farm, University of Saskatchewan in 2016 and 2017. Fababean ( Vicia faba L.) crop tolerance to nine herbicide tank mixtures was evaluated with 2 rates distributed in a randomized complete block design (RCBD) with 4 blocks. The trial was imaged using a multispectral camera with a ground sample distance (GSD) of 1.2 cm, one week after the treatment application. Visual ratings of growth reduction and physiological chlorosis were recorded simultaneously with imaging. The optimized soil-adjusted vegetation index (OSAVI) was calculated from the thresholded orthomosaics. The UAV-based vegetation index (OSAVI) produced more precise results compared to visual ratings for both years. The coefficient of variation (CV) of OSAVI was ~1% when compared to 18-43% for the visual ratings. Furthermore, Tukey's honestly significance difference (HSD) test yielded a more precise mean separation for the UAV-based vegetation index than visual ratings. The significant correlations between OSAVI and the visual ratings from the study suggest that undesirable variability associated with visual assessments can be minimized with the UAV-based approach. UAV-based imagery methods had greater precision than the visual-based ratings for crop herbicide damage. These methods have the potential to replace visual ratings and aid in screening crops for herbicide tolerance.
Background Soil moisture deficiency causes yield reduction and instability in faba bean ( Vicia faba L.) production. The extent of sensitivity to drought stress varies across accessions originating from diverse moisture regimes of the world. Hence, we conducted successive greenhouse experiments in pots and rhizotrons to explore diversity in root responses to soil water deficit. Methods A set of 89 accessions from wet and dry growing regions of the world was defined according to the Focused Identification of Germplasm Strategy and screened in a perlite-sand medium under well watered conditions in a greenhouse experiment. Stomatal conductance, canopy temperature, chlorophyll concentration, and root and shoot dry weights were recorded during the fifth week of growth. Eight accessions representing the range of responses were selected for further investigation. Starting five days after germination, they were subjected to a root phenotyping experiment using the automated phenotyping platform GROWSCREEN-Rhizo. The rhizotrons were filled with peat-soil under well watered and water limited conditions. Root architectural traits were recorded five, 12, and 19 days after the treatment (DAT) began. Results In the germplasm survey, accessions from dry regions showed significantly higher values of chlorophyll concentration, shoot and root dry weights than those from wet regions. Root and shoot dry weight as well as seed weight, and chlorophyll concentration were positively correlated with each other. Accession DS70622 combined higher values of root and shoot dry weight than the rest. The experiment in GROWSCREEN-Rhizo showed large differences in root response to water deficit. The accession by treatment interactions in taproot and second order lateral root lengths were significant at 12 and 19 DAT, and the taproot length was reduced up to 57% by drought. The longest and deepest root systems under both treatment conditions were recorded by DS70622 and DS11320, and total root length of DS70622 was three times longer than that of WS99501, the shortest rooted accession. The maximum horizontal distribution of a root system and root surface coverage were positively correlated with taproot and total root lengths and root system depth. DS70622 and WS99501 combined maximum and minimum values of these traits, respectively. Thus, roots of DS70622 and DS11320, from dry regions, showed drought-avoidance characteristics whereas those of WS99501 and Mèlodie/2, from wet regions, showed the opposite. Discussion The combination of the germplasm survey and use of GROWSCREEN-Rhizo allowed exploring of adaptive traits and detection of root phenotypic markers for potential drought avoidance. The greater root system depth and root surface coverage, exemplified by DS70622 and DS11320, can now be tested as new sources of drought tolerance.
The manuscript presents a procedure for optimal sample preparation and the mapping of the spatial distribution of metal ions and nanoparticles in plant roots using laser-induced breakdown spectroscopy (LIBS) in a double-pulse configuration (DP LIBS) in orthogonal reheating mode. Two Nd:YAG lasers were used; the first one was an ablation laser (UP-266 MACRO, New Wave, USA) with a wavelength of 266nm, and the second one (Brilliant, Quantel, France), with a fundamental wavelength of 1064nm, was used to reheat the microplasma. Seedlings of Vicia faba were cultivated for 7 days in CuSO 4 or AgNO 3 solutions with a concentration of 10µmoll -1 or in a solution of silver nanoparticles (AgNPs) with a concentration of 10µmoll -1 of total Ag, and in distilled water as a control. The total contents of the examined metals in the roots after sample mineralization as well as changes in the concentrations of the metals in the cultivation solutions were monitored by ICP-OES. Root samples embedded in the TissueTek medium and cut into 40µm thick cross sections using the Cryo-Cut Microtome proved to be best suited for an accurate LIBS analysis with a 50µm spatial resolution. 2D raster maps of elemental distribution were created for the emission lines of Cu(I) at 324.754nm and Ag(I) at 328.068nm. The limits of detection of DP LIBS for the root cross sections were estimated to be 4pg for Cu, 18pg for Ag, and 3pg for AgNPs. The results of Ag spatial distribution mapping indicated that unlike Ag + ions, AgNPs do not penetrate into the inner tissues of Vicia faba roots but stay in their outermost layers. The content of Ag in roots cultivated in the AgNP solution was one order of magnitude lower compared to roots cultivated in the metal ion solutions. The significantly smaller concentration of Ag in root tissues cultivated in the AgNP solution also supports the conclusion that the absorption and uptake of AgNPs by roots of Vicia faba is very slow. LIBS mapping of root sections represents a fast analytical method with sufficient precision and spatial resolution that can provide very important information for researchers, particularly in the fields of plant science and ecotoxicology.
The standard methodology for assessing seed size distribution of pulses is to sieve seeds into size classes, weighing each class and calculating a weighted mean of the sieve sizes (seed size index). A single unit measure of uniformity of size within a sample via sieving is not available, despite being an important trait in terms of appearance, ease of milling, and consistency in processing. This study investigated several different models for estimating the size variability within seed samples by using a variety of desi and kabuli chickpea, faba bean, lupin, lentil, and mungbean samples. Fitting a normal distribution to the frequency distribution and using the estimated variance parameter as a measure for seed size variability was found to be the most suitable method for attaining seed size and a single value of uniformity. Mean seed size (SSₙₒᵣₘ) and within‐sample size variability (SVₙₒᵣₘ) were unrelated variables, thus allowing selection of more uniform samples of any desired size in breeding programs. Sieve selection is discussed and needs to be appropriate for the samples under investigation. Examples of using the R software to calculate these measures and the R functions needed are available as supplementary files to facilitate the use of this proposed method. This method will be useful to pulse breeders and researchers and in the development of image analysis methods characterizing seed sizes of pulse samples and other grains.
The understanding of crop domestication is dependent on tracking the original geographical distribution of wild relatives. The faba bean (Vicia faba L.) is economically important in many countries around the world; nevertheless, its origin has been debated because its ancestor could not be securely identified. Recent investigations in the site of el-Wad (Mount Carmel, Israel), provide the first and, so far, only remains of the lost ancestor of faba bean. X-ray CT scan analysis of the faba beans provides the first set of measurements of the biometry of this species before its domestication. The presence of wild specimens in Mount Carmel, 14,000 years ago, supports that the wild variety grew nearby in the Lower Galilee where the first domestication was documented for Neolithic farmers 10,200 years ago.