Weeds compete with crops for vital resources, reducing yield and quality. Although recent deep learning advances enable precise weed detection, variability in species, growth stages, and densities presents considerable challenges. To address these challenges, this study proposes a novel weed detection approach that combines two independent models. The first model employs an image segmentation network that identifies crops and removes them from the images. The remaining image is divided into grid regions, which are analyzed by a second image classification model to differentiate weeds from background soil. An improved Lightweight Efficient Segmentation Network based on You Only Look Once version 8 (LESNet-YOLOv8), was designed and developed to achieve efficient and accurate cabbage segmentation. Key enhancements include replacing conventional convolution operations with more efficient alternatives, optimizing feature extraction and fusion processes, and incorporating an optimized detection head. The improved network model is compact and only 2.9 megabytes (MB) in size. Additionally, the average accuracy at the Intersection over Union (IoU) threshold of 0.5 on the mask reaches 98.2 % in the segmentation task. Moreover, the proposed method yielded F1 (the harmonic mean of precision and recall) scores of 91.4 % and 92.4 % for weed and non-weed regions, respectively. The method proposed in this study offers a deployment-oriented, practical, and scalable solution that directly links grid-based detection with nozzle-level actuation, bridging computer vision and precision weed management in real-world agricultural applications.

Effective weed management in imidazolinone (IMI)-resistant corn (Zea mays L.) and soybean (Glycine max (L.) Merr.) strip intercropping systems is critical for maximizing productivity and ensuring long-term sustainability. In regions where genetically modified crops are not permitted, IMI herbicides are the only choice for postemergence (POST) grass weed control in this intercropping system, raising concerns about overuse and herbicide resistance. This study evaluated weed control efficacy and crop safety of tank mixtures combining IMI herbicides with soil-applied residual herbicides. The results showed that preemergence (PRE) applications of IMI herbicides alone, imazamox, imazapic, or imazethapyr, provided inconsistent control of barnyardgrass (Echinochloa crus-galli (L.) P. Beauv.), perilla mint (Perilla frutescens (L.) Britton), and redroot pigweed (Amaranthus retroflexus L.). In contrast, PRE tank mixtures of IMI herbicides with residual herbicides demonstrated effective control. PRE application of imazamox + S-metolachlor achieved 91%, 95%, and 100% control of barnyardgrass, perilla mint, and redroot pigweed, respectively, while imazethapyr + S-metolachlor provided 100%, 98%, and 100% control of the same species. POST applications of IMI herbicides combined with either pendimethalin or S-metolachlor resulted in >90% control of barnyardgrass and common purslane (Portulaca oleracea L.) by 4 weeks after treatment (WAT). Field experiments confirmed that the imazamox + pendimethalin mixture not only controlled existing weeds but also extended control duration and outperformed the individual herbicides. By 6 WAT, imazamox alone provided ≤77% control and pendimethalin provided ≤66% control of common purslane, redroot pigweed, and smooth crabgrass (Digitaria ischaemum (Schreb.) Schreb. ex Muhl.). In contrast, their combination delivered >90% control of all four weed species and was crop-safe, with no observed injury or yield reduction in either corn or soybean. By integrating multiple modes of action and extending residual weed suppression, these mixtures may reduce dependence on repeated IMI-only POST programs while maintaining crop safety. Overall, integrating IMI herbicides with residual products provides an effective strategy for weed management in IMI-resistant corn–soybean strip intercropping systems.

Global climate change is profoundly reshaping the competitive landscape between C4 weeds and crops. This review focuses on the unique CO2-concentrating mechanism of C4 weeds, elucidating the physiological and biochemical basis for their photosynthetic advantage under high-temperature and drought conditions, and systematically explaining how this advantage drives the restructuring of weed communities and the expansion of their geographical distribution. On this basis, it analyzes the regulatory pathways through which elevated CO2 concentration, temperature changes, and water stress affect the photosynthetic efficiency of C4 weeds, as well as the resulting shifts in aboveground and belowground resource competition patterns. In response to the limitations of current control strategies, a multidimensional management framework integrating agronomic, chemical, and biotechnological approaches is proposed. Finally, key directions for future research are identified, including the molecular dissection of the C4 photosynthetic regulatory network, the development of predictive models for climate-weed-crop interactions, and the development of targeted control technologies, thereby providing a forward-looking framework for addressing climate-driven weed risks.

Weed control is crucial for optimizing corn yield. In recent years, advances in computer vision and deep learning have created new opportunities for precision agriculture. However, annotating weed datasets is typically time-consuming, labor-intensive, and costly. To address this challenge, this study proposes an indirect weed detection strategy that reduces reliance on explicit weed annotations by focusing on accurate crop segmentation. Specifically, we develop YOLO-CornSeg, a lightweight segmentation model based on an improved YOLOv8n architecture, designed for precise corn seedling segmentation. The model incorporates a C2f_DWR module to enhance multi-scale feature extraction and a Segment_Efficient head to improve segmentation performance while maintaining computational efficiency. Based on the resulting segmentation masks, an indirect weed detection strategy is applied, in which non-crop green regions are identified as weeds using HSV-based image processing. Experimental results show that YOLO-CornSeg achieves a mean Intersection over Union (mIoU) of 91.1% with a model size of 8.3 MB, outperforming several state-of-the-art two-stage semantic segmentation models while maintaining low computational complexity and a compact model size. The improved segmentation accuracy further enhances the reliability of downstream weed inference. Overall, this study highlights the potential of combining lightweight crop segmentation with indirect weed detection strategies to support precision herbicide application.

Phosphatidic acid (PA), a key signaling molecule in plant immunity, activates MAPK-mediated and reactive oxygen species (ROS)-mediated defense responses. Exogenous application of PA significantly reduces Potato virus Y (PVY) genomic RNA accumulation and alleviates disease symptoms in Nicotiana benthamiana, highlighting its potential as an immune elicitor for crop disease management. This chapter provides detailed protocols for PA solution preparation, MAPK pathway activation analysis (by Western blot and qPCR), and virus quantification to guide PA application for broader plant virus control in crops.

Abstract Andrew Otis Jackson (1941–2025) was a world-renowned plant virologist who made pioneering contributions to the study of both plant positive- and negative-strand RNA viruses. Since his first visit to China in 1998, Professor Jackson forged deep and lasting connections with the Chinese plant virology community. Through his dedicated teaching and close research collaborations, he generously shared his scientific insights, methodologies, and international perspectives with students and young scholars at several Chinese universities, most notably China Agricultural University and Zhejiang University, significantly advancing plant virology research in China and fostering the career development of young scientists. This article, rooted in personal experiences and shared memories, seeks to honor his enduring scientific legacy, recount the journey of his international collaboration with China, commemorate his profound contributions, and express the deep gratitude felt by those who had the privilege of crossing paths with him.

Palmer amaranth (Amaranthus palmeri S. Watson), native to North America, is one of the most prominent invasive weed species on agricultural land. Acetolactate synthase (ALS)-resistant A. palmeri (Amaranthus palmeri) is widespread, while the research focus on resistance pattern and molecular basis of A. palmeri to imazethapyr is seldom documented in China. An A. palmeri population that survived the recommended rate of imazethapyr was collected in Shandong Province, China. The resistant mechanism and pattern of A. palmeri to imazethapyr was investigated. Dose–response assay showed that the resistant (R) population displayed a high resistance level (292.5-fold) to imazethapyr compared with the susceptible (S) population. Sequence analysis of the ALS gene revealed that nucleotide mutations resulted in three resistance-conferring amino acid substitutions, Pro-197-Ile, Trp-574-Leu, and Ser-653-Asp, in the individual plants of the R population. An in vitro enzyme assay indicated that the ALS was relatively unsusceptible to imazethapyr in the R population, showing a resistance index of 88.6-fold. ALS gene expression and copy number did not confer resistance to imazethapyr in the R population. Pro-197-Ile is the first reported amino acid substitution conferring ALS resistance to A. palmeri. This is the first case of an imazethapyr-resistant A. palmeri biotype in China.

BACKGROUND:Precision weed mapping in turf according to its susceptibility to selective herbicides allows the smart sprayer to spot-spray the most pertinent herbicides onto the susceptible weeds. The objective of this study was to evaluate the feasibility of implementing herbicide susceptibility-based weed mapping using deep convolutional neural networks (DCNNs) to facilitate targeted and efficient herbicide applications. Additionally, applying path-planning algorithms to weed mapping data to guide the spraying nozzle ensures minimal travel paths for herbicide application. RESULTS:DenseNet achieved high precision, recall, overall accuracy, and F1 score values for all categories of herbicides and no herbicides, with F1 scores ranging from 0.996 to 0.999 in the validation dataset and from 0.992 to 0.997 in the testing dataset. The average accuracies attained by DenseNet, GoogLeNet and ResNet were 0.9985, 0.9953 and 0.9980, respectively. By considering both accuracy and computational efficiency, the ResNet model was identified as the most effective among the models compared to weed detection. The performance of the Christofides, Greedy and 2-opt algorithms in optimizing path planning for single or dual spraying nozzles was compared and analyzed. The Greedy algorithm proved the most efficient in optimizing the nozzle's trajectory. CONCLUSION:Implementing herbicide susceptibility-based weed mapping facilitates targeted herbicide application by directing the nozzle to the grid cells containing the weeds susceptible to the herbicides. Moreover, the strategic integration of herbicide susceptibility-based weed mapping with optimized path planning for the spraying mechanism can be adeptly implemented on smart sprayers, which could effectively reduce the herbicide input. © 2025 Society of Chemical Industry.

This paper addresses deficiencies in existing spray carts and suspended sprayers regarding operational scenarios, spray coverage, versatility, and wall film thickness adjustment by designing a rail-mounted rotating nozzle application robot. Static analysis of the robot frame verifies compliance with strength and stiffness requirements. Motor torque calculations ensure stable and reliable nozzle rotation. Geometric modeling derives optimal link parameters for automated nozzle angle control. ANSYS Fluent simulations characterize static spray coverage, analyzing quantitative relationships between nozzle height, angle, and spray distance. SolidWorks Motion establishes a coupled model of nozzle rotation and cart translation to obtain spray trajectories under varying speeds. Coupled Fluent simulations further evaluate wall film thickness distribution patterns under dynamic spraying conditions. The findings provide a theoretical foundation and technical reference for structural optimization and precise spraying control in greenhouse spraying robot systems.

Effective weed management is essential for maintaining stable agricultural productivity and ensuring crop yield. Recent advancements in computer vision technologies have significantly enhanced weed detection efficiency. However, the diversity of weed species and irregular spatial distribution present considerable challenges to traditional weed detection methods, which often exhibit low accuracy in practical applications. Additionally, the construction of large-scale, high-quality datasets encompassing a wide variety of weed species is constrained by high costs and time limitations. Indirect weed recognition methods offer a promising solution to this issue. This approach first employs an image segmentation network to precisely identify and remove crop regions from the original image, followed by an image processing algorithm that extracts green pixels for weed identification. Based on the indirect weed detection method, this study optimizes the YOLOv8n-seg architecture to improve segmentation accuracy while maintaining model efficiency, leading to the development of the enhanced ESL-YOLOv8 segmentation model. To validate the effectiveness of the proposed method, a maize-specific segmentation dataset was constructed, and extensive experiments were conducted. The results show that the improved ESL-YOLOv8 model achieves a mask mAP50 value of 94.1% based on the mask, which is superior to the original YOLOv8nseg(92.0%). Furthermore, the model size is reduced to 5.7 MB, compared to 6.8 MB for the original model, with a decrease in computational complexity. These findings confirm that the proposed framework provides a lightweight, high-precision, and practical solution for weed detection in agricultural fields, exhibiting enhanced robustness and significant engineering applicability.

Numerous studies suggest that virus infections can improve drought resistance in host plants, but the underlying mechanisms remain unclear. In this study, we used Turnip mosaic virus (TuMV), Potato virus Y (PVY), Tomato bushy stunt virus (TBSV) and the host Nicotiana benthamiana as model systems to investigate these mechanisms. Our findings reveal that the abscisic acid (ABA) signalling pathway is strongly and durably induced by +RNA virus infection and is essential for virus-induced drought tolerance. Notably, although ABA content increased following virus infection, this elevation was not necessary for downstream ABA signalling or virus-induced drought tolerance. Instead, +RNA virus-induced drought tolerance relies on N. benthamiana phospholipase Dα1 (NbPLDα1)-derived phosphatidic acid (PA). Knockout of NbPLDα1 or disruption of the interaction between viral proteins and NbPLDα1 impaired the ability of +RNA viruses to activate ABA signalling and enhance drought tolerance. The virus-induced increase in ABA levels appears to result from feedback regulation by PA-activated ABA signalling. Overall, our results suggest that +RNA viruses improve plant drought tolerance by modulating NbPLDα1-derived PA rather than by promoting ABA production.

Computer vision-based precision weed control has proven effective in reducing herbicide usage, lowering weed management costs, and enhancing sustainability in modern agriculture. However, developing deep learning models remains challenging due to the effort required for weed dataset annotation and the difficulty of identifying weeds at different stages and densities in complex field conditions. To address these challenges, this study introduces an indirect weed detection method that combines deep learning and image processing techniques. The proposed approach first employs an object detection network to identify and label crops within the images. Subsequently, image processing techniques are applied to segment the remaining green pixels, thereby enabling indirect detection of weeds. Furthermore, a novel detection network-CD-YOLOv10n (You Only Look Once version 10 nano)-was developed based on the YOLOv10 framework to optimize computational efficiency. Redesigning the backbone (C2f-DBB) and integrating an optimized upsampling module (DySample) permitted the network to achieve higher detection accuracy while maintaining a lightweight structure. Specifically, the model achieved a mean average precision (mAP50) of 98.1%, which is a 1.4% percentage-point increase compared with the YOLOv10n baseline, a relevant improvement given the already strong baseline performance. At the same time, compared with YOLOv10n, its GFLOPs (giga floating-point operations per second) were reduced by 22.62%, and the number of parameters decreased by 15.87%. These innovations make CD-YOLOv10n highly suitable for deployment on resource-constrained platforms.

Automatic precision herbicide application offers significant potential for reducing herbicide use in turfgrass weed management. However, developing accurate and reliable neural network models is crucial for achieving optimal precision weed control. The reported neural network models in previous research have been limited by specific geographic regions, weed species, and turfgrass management practices, restricting their broader applicability. The objective of this research was to evaluate the feasibility of deploying a single, robust model for weed classification across a diverse range of weed species, considering variations in species, ecotypes, densities, and growth stages in bermudagrass turfgrass systems across different regions in both China and the United States. Among the models tested, ResNeXt152 emerged as the top performer, demonstrating strong weed detection capabilities across 24 geographic locations and effectively identifying 14 weed species under varied conditions. Notably, the ResNeXt152 model achieved an F 1 score and recall exceeding 0.99 across multiple testing scenarios, with a Matthews correlation coefficient (MCC) value surpassing 0.98, indicating its high effectiveness and reliability. These findings suggest that a single neural network model can reliably detect a wide range of weed species in diverse turf regimes, significantly reducing the costs associated with model training and confirming the feasibility of using one model for precision weed control across different turf settings and broad geographic regions.

Accurate weed detection in crop fields remains a challenging task due to the diversity of weed species and their visual similarity to crops, especially under natural field conditions where lighting and occlusion vary. Traditional methods typically attempt to directly identify various weed species, which demand large-scale, finely annotated datasets and often suffer from low generalization. To address these challenges, this study proposes a novel dual-model framework that simplifies the task by dividing it into two tractable stages. First, a crop segmentation network is used to identify and remove cabbage (Brassica oleracea L. ssp. pekinensis) regions from field images. Since crop categories are visually consistent and singular, this stage achieves high precision with relatively low complexity. The remaining non-crop areas, which contain only weeds and background, are then subdivided into grid cells. Each cell is classified by a second lightweight classification network as either background, broadleaf weeds, or grass weeds. The classification model achieved F1 scores of 95.1%, 91.1%, and 92.2% for background, broadleaf weeds, and grass weeds, respectively. This two-stage approach transforms a complex multi-class detection task into simpler, more manageable subtasks, improving detection accuracy while reducing annotation burden and enhancing robustness under the tested field conditions.

BACKGROUND:Differentiating between potato common scab, powdery scab, and the physiological disorder of enlarged corky lenticels is challenging due to their similar visual symptoms. To address this, we propose YOLOv8-ST, an enhanced deep learning model that incorporates the Swin Transformer and Triplet Attention modules to effectively distinguish between these visually similar tuber blemishes. RESULTS:YOLOv8-ST is an enhanced YOLOv8 model with the integration of Triplet Attention and the Swin Transformer, which achieved significant accuracy improvements. Compared to the baseline of YOLOv3, YOLOv5, YOLOv6, and YOLOv8, YOLOv8-ST achieved the highest precision (0.903), recall (0.831), F1-score (0.866), mAP@0.5 (0.931), and mAP@0.5:0.95 (0.616), with strong performance in detecting common scab and powdery scab (both >0.9 at mAP@0.5 or precision). Detection outputs showed higher confidence (e.g., 0.94 for scab), fewer false positives, and no missed lesions, outperforming models prone to misclassification or overlap. CONCLUSION:The YOLOv8-ST model enables fast, accurate, and reliable detection of common scab, powdery scab, and enlarged lenticels on potato tubers. This field-deployable solution supports early disease diagnosis and timely intervention, thus reducing crop losses. The model is available through the mobile app Plant Guardian, enabling growers to identify potato skin blemishes directly in the field, thereby advancing both practical disease management and agricultural AI applications. © 2025 Society of Chemical Industry.

The strip intercropping system is one of the important production strategy for maize and soybean. However, it faces significant yield losses due to the lack of safe and effective postemergence herbicides compatible with both crops. In this context, the introduction of imidazolinone-tolerant maize presents a potential solution, yet the associated crop traits, yield outcomes, herbicidal efficacy, and economic impacts have not been thoroughly evaluated. Therefore, this study systematically compares two weed control strategies in maize-soybean strip intercropping system: non-segregated weeding (NSW), which uses imidazolinone-tolerant maize to allow for shared herbicide application, and segregated weeding (SW), which employs a dual-system sprayer for separate herbicide treatments for maize and soybean. A control group with no weed control (NW) was also included to compare the results across treatments. Our results revealed that the differences in plant height and stem diameter between maize and soybean were not significant between NSW and SW, though both were substantially lower compared to their respective monocultures. Compared to SW, the NSW treatment increased the leaf area index, total dry matter accumulation, and grain yield of soybean by 33%, 17%, and 79%, respectively. For maize, these parameters were marginally higher but not significant, indicating that the NSW treatment benefited soybean growth and yield more than maize in maize-soybean strip intercropping system. Overall, maize and soybean under SW and NSW achieved land equivalent ratios of 0.96 and 0.99 for maize, and 0.40 and 0.80 for soybean, respectively, suggesting that the NSW strategy provided better weed control and allowed for more efficient land use for both maize and soybean. Specifically, in terms of weed suppression, NSW outperformed SW, with the number and fresh weight of weeds (Gramineae, Broadleaf, Cyperaceous) reduced to 16% and 20% of those in SW, and to 5% and 4% of those in NW, respectively. Moreover, NSW increased weeding speed by fivefold, reduced herbicide and spraying costs by $37.05 USD ha-1, and enhanced net benefits by 58%, reaching $3414.12 USD ha-1. These findings demonstrate that NSW, based on imidazolinone-tolerant maize, offers a more convenient, economical, and efficient weed management strategy for maize-soybean strip intercropping system.

Long-term use of naproxen can lead to serious side effects. Inspired by the biological activity of cinnamic acid, a series of cinnamic acid derivatives containing naproxen were designed and synthesized, and their anti-inflammatory activities and mechanisms were explored in vitro. Our results indicated that all of naproxen derivatives showed more significant inhibition against lipopolysaccharide (LPS)-induced nitric oxide (NO) production and had a lower degree of cytotoxicity than that of naproxen. The present studies revealed that compound 23 (IC50 = 5.66 ± 1.66 µM) markedly inhibited the LPS-induced NO production and the over-expression of pro-inflammatory cytokines, including interleukin (IL)-1β, inducible NO synthase (iNOS), and cyclooxygenase-2 (COX-2). Furthermore, it blocked the activation of NF-κB signaling pathway and pyrin domain-containing protein 3 (NLRP-3) inflammasome in a concentration-dependent manner. Additionally, docking studies confirmed that compound 23 exhibited a well-fitting into the NLRP3 active site. Considering these results, compound 23 might be a novel NLRP3 inhibitor to treat inflammatory diseases.

In small-plot experiments, weed scientists have traditionally estimated herbicide efficacy through visual assessments or manual counts with wooden frames—methods that are time-consuming, labor-intensive, and error-prone. This study introduces a novel mobile application (app) powered by convolutional neural networks (CNNs) to automate the evaluation of weed coverage in turfgrass. The mobile app automatically segments input images into 10 by 10 grid cells. A comparative analysis of EfficientNet, MobileNetV3, MobileOne, ResNet, ResNeXt, ShuffleNetV1, and ShuffleNetV2 was conducted to identify weed-infested grid cells and calculate weed coverage in bahiagrass ( Paspalum notatum Flueggé), dormant bermudagrass [ Cynodon dactylon (L.) Pers.], and perennial ryegrass ( Lolium perenne L.). Results showed that EfficientNet and MobileOne outperformed other models in detecting weeds growing in bahiagrass, achieving an F 1 score of 0.988. For dormant bermudagrass, ResNet performed best, with an F 1 score of 0.996. Additionally, app-based coverage estimates (11%) were highly consistent with manual assessments (11%), showing no significant difference (P = 0.3560). Similarly, ResNeXt achieved the highest F 1 score of 0.996 for detecting weeds growing in perennial ryegrass, with app-based and manual coverage estimates also closely aligned at 10% (P = 0.1340). High F 1 scores across all turfgrass types demonstrate the models’ ability to accurately replicate manual assessments, which is essential for herbicide efficacy trials requiring precise weed coverage data. Moreover, the time for weed assessment was compared, revealing that manual counting with 10 by 10 wooden frames took an average of 39.25, 37.25, and 42.25 s per instance for bahiagrass, dormant bermudagrass, and perennial ryegrass, respectively, whereas the app-based approach reduced the assessment times to 8.23, 7.75, and 14.96 s, respectively. These results highlight the potential of deep learning–based mobile tools for fast, accurate, scalable weed coverage assessments, enabling efficient herbicide trials and offering labor and cost savings for researchers and turfgrass managers.

Computer vision-based precision spraying of herbicides presents a promising avenue for reducing herbicide input and weed control costs. Nonetheless, weed detection in wheat (Triticum aestivum L.) remains challenging. Developing an effective and reliable neural network for weed detection requires substantial labeled data for training. However, labeling data is time-consuming and labor-intensive. To address this challenge, the present study introduces semi-supervised learning (SSL) into the domain of weed detection in wheat. The performance of four SSL methods was thoroughly evaluated and compared with that of a fully supervised learning (FSL) method on a dataset with a limited amount of labeled images. Experimental results showed that the Fixmatch method, an SSL approach, outperformed the FSL method, exhibiting significantly higher accuracy (ACC) with a limited number of labeled images. The ACC of Fixmatch was 85.4%, which was 7.3% higher than the FSL method. In further analysis, the performance of models trained on a dataset containing 100, 200, 300, 400, 500, or 1000 labeled images per class was tested. Compared with FSL, SSL achieved the greatest improvement when the number of labels was 200. At the same time, Fixmatch achieved satisfactory performance, ACC, recall, and precision reached 94.8%, 94.8%, and 95.2%, respectively, and the F1 score was 95%. In summary, these results suggest that using the SSL method could yield a high-performing model when training with a limited number of labeled images, requiring less training costs and lower demands on manpower.

BACKGROUNDMachine vision-based precision weed management is a promising solution to substantially reduce herbicide input and weed control cost. The objective of this research was to compare two different deep learning-based approaches for detecting weeds in cabbage: (1) detecting weeds directly, and (2) detecting crops by generating the bounding boxes covering the crops and any green pixels outside the bounding boxes were deemed as weeds.RESULTSThe precision, recall, F1-score, mAP0.5, mAP0.5:0.95 of You Only Look Once (YOLO) v5 for detecting cabbage were 0.986, 0.979, 0.982, 0.995, and 0.851, respectively, while these metrics were 0.973, 0.985, 0.979, 0.993, and 0.906 for YOLOv8, respectively. However, none of these metrics exceeded 0.891 when detecting weeds. The reduced performances for directly detecting weeds could be attributed to the diverse weed species at varying densities and growth stages with different plant morphologies. A segmentation procedure demonstrated its effectiveness for extracting weeds outside the bounding boxes covering the crops, and thereby realizing effective indirect weed detection.CONCLUSIONThe indirect weed detection approach demands less manpower as the need for constructing a large training dataset containing a variety of weed species is unnecessary. However, in a certain case, weeds are likely to remain undetected due to their growth in close proximity with crops and being situated within the predicted bounding boxes that encompass the crops. The models generated in this research can be used in conjunction with the machine vision subsystem of a smart sprayer or mechanical weeder. (c) 2024 Society of Chemical Industry. The presence of various weed species increased the complexity of feature extraction for direct weed detection. An indirect weed detection, achieved by extracting green pixels outside the predicted bounding boxes covering the cabbage; however, weeds remained undetected when they grow in close proximity to the cabbage. image