Search

BACKGROUND: Machine 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. RESULTS: The 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. CONCLUSION: The 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. © 2024 Society of Chemical Industry. [ABSTRACT FROM AUTHOR]

Automated precision weed control requires visual methods to discriminate between crops and weeds. State-of-the-art plant detection methods fail to reliably detect weeds, especially in dense and occluded scenes. In the past, using hand-crafted detection models, both color (RGB) and depth (D) data were used for plant detection in dense scenes. Remarkably, the combination of color and depth data is not widely used in current deep learning-based vision systems in agriculture. Therefore, we collected an RGB-D dataset using a stereo vision camera. The dataset contains sugar beet crops in multiple growth stages with a varying weed densities. This dataset was made publicly available and was used to evaluate two novel plant detection models, the D-model, using the depth data as the input, and the CD-model, using both the color and depth data as inputs. For ease of use, for existing 2D deep learning architectures, the depth data were transformed into a 2D image using color encoding. As a reference model, the C-model, which uses only color data as the input, was included. The limited availability of suitable training data for depth images demands the use of data augmentation and transfer learning. Using our three detection models, we studied the effectiveness of data augmentation and transfer learning for depth data transformed to 2D images. It was found that geometric data augmentation and transfer learning were equally effective for both the reference model and the novel models using the depth data. This demonstrates that combining color-encoded depth data with geometric data augmentation and transfer learning can improve the RGB-D detection model. However, when testing our detection models on the use case of volunteer potato detection in sugar beet farming, it was found that the addition of depth data did not improve plant detection at high vegetation densities. [ABSTRACT FROM AUTHOR]

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. • This paper introduces semi-supervised learning to detect weeds growing in wheat. • Semi-supervised learning outperformed fully supervised learning for weed detection. • Using semi-supervised methods reduced the requirement of training costs with fewer labeled data. [ABSTRACT FROM AUTHOR]

Convolutional neural networks (CNNs) have revolutionized the weed detection process with tremendous improvements in precision and accuracy. However, training these models is time-consuming and computationally demanding; thus, training weed detection models for every crop-weed environment may not be feasible. It is imperative to evaluate how a CNN-based weed detection model trained for a specific crop may perform in other crops. In this study, a CNN model was trained to detect morningglories and grasses in cotton. Assessments were made to gauge the potential of the very model in detecting the same weed species in soybean and corn under two levels of detection complexity (levels 1 and 2). Two popular object detection frameworks, YOLOv4 and Faster R-CNN, were trained to detect weeds under two schemes: Detect_Weed (detecting at weed/crop level) and Detect_Species (detecting at weed species level). In addition, the main cotton dataset was supplemented with different amounts of non-cotton crop images to see if cross-crop applicability can be improved. Both frameworks achieved reasonably high accuracy levels for the cotton test datasets under both schemes (Average Precision-AP: 0.83–0.88 and Mean Average Precision-mAP: 0.65–0.79). The same models performed differently over other crops under both frameworks (AP: 0.33–0.83 and mAP: 0.40–0.85). In particular, relatively higher accuracies were observed for soybean than for corn, and also for complexity level 1 than for level 2. Significant improvements in cross-crop applicability were further observed when additional corn and soybean images were added to the model training. These findings provide valuable insights into improving global applicability of weed detection models. [ABSTRACT FROM AUTHOR]

This review is designed to address various alternative weed-control practices and their possibilities in the fruit orchard in terms of sustainability. Correct weed management and maintenance of adequate orchard biodiversity are crucial for sustainable orchard soil management. The key is to practice an alternative weed-management approach (single or integrated) rather than to use possibly harmful chemicals only. Integration of modern equipment with a shallow tillage system can provide effective weed control in tree rows, including optimised tree performance and soil biodiversity. Living mulch suppresses weeds and enhances orchard biodiversity, while selection of less competitive and less pest-attracting species is crucial. Plastic covers offer long-term weed control, but additional nutrient amendments are required to maintain the balanced fertility of the soil. Wood chip mulch is suggested where the materials are available on or near the farm, and where there is lower incidence of perennial weeds. High pressure water and robotic systems are still in their infancy for fruit orchards, and required more research to confirm their efficiency.

Convolutional neural networks (CNNs) have revolutionized the weed detection process with tremendous improvements in precision and accuracy. However, training these models is time-consuming and computationally demanding; thus, training weed detection models for every crop-weed environment may not be feasible. It is imperative to evaluate how a CNN-based weed detection model trained for a specific crop may perform in other crops. In this study, a CNN model was trained to detect morningglories and grasses in cotton. Assessments were made to gauge the potential of the very model in detecting the same weed species in soybean and corn under two levels of detection complexity (levels 1 and 2). Two popular object detection frameworks, YOLOv4 and Faster R-CNN, were trained to detect weeds under two schemes: Detect_Weed (detecting at weed/crop level) and Detect_Species (detecting at weed species level). In addition, the main cotton dataset was supplemented with different amounts of non-cotton crop images to see if cross-crop applicability can be improved. Both frameworks achieved reasonably high accuracy levels for the cotton test datasets under both schemes (Average Precision-AP: 0.83–0.88 and Mean Average Precision-mAP: 0.65–0.79). The same models performed differently over other crops under both frameworks (AP: 0.33–0.83 and mAP: 0.40–0.85). In particular, relatively higher accuracies were observed for soybean than for corn, and also for complexity level 1 than for level 2. Significant improvements in cross-crop applicability were further observed when additional corn and soybean images were added to the model training. These findings provide valuable insights into improving global applicability of weed detection models.

Remote sensing using unmanned aerial vehicles (UAVs) for weed detection is a valuable asset in agriculture and is vastly used for site-specific weed control. Alongside site-specific methods, time-specific weed control is another critical aspect of precision weed control where, by using different models, it is possible to determine the time of weed species emergence. This study combined site-specific and time-specific weed control methods to explore their collective benefits for precision weed control. Using the AlertInf model, a weed emergence prediction model, the cumulative emergence of Sorghum halepense was calculated, following the selection of the best date for the UAV survey when the emergence was predicted to be at 96%. The survey was executed using a UAV with visible range sensors, resulting in an orthophoto with a resolution of 3 cm, allowing for good weed detection. The orthophoto was post-processed using two separate methods: an artificial neural network (ANN) and the visible atmospherically resistant index (VARI) to discriminate between the weeds, the crop, and the soil. Finally, a model was applied for the creation of prescription maps with different cell sizes (0.25 m², 2 m², and 3 m²) and with three different decision-making thresholds based on pixels identified as weeds (>1%, >5%, and >10%). Additionally, the potential savings in herbicide use were assessed using two herbicides (Equip and Titus Mais Extra) as examples. The results show that both classification methods have a high overall accuracy of 98.6% for ANN and 98.1% for VARI, with the ANN having much better results concerning user/producer accuracy and Cohen's Kappa value (k=83.7 ANN and k=72 VARI). The reduction percentage of the area to be sprayed ranged from 65.29% to 93.35% using VARI and from 42.43% to 87.82% using ANN. The potential reduction in herbicide use was found to be dependent on the area. For the Equip herbicide, this reduction ranged from 1.32 L/ha to 0.28 L/ha for the ANN; with VARI the reduction in the amounts used ranged from 0.80 L/ha to 0.15 L/ha. Meanwhile, for Titus Mais Extra herbicide, the reduction ranged from 46.06 g/ha to 8.19 g/ha in amounts used with the ANN; with VARI the amount reduction ranged from 27.77 g/ha to 5.32 g/ha. These preliminary results indicate that combining site-specific and timespecific weed control might significantly reduce herbicide use with direct benefits for the environment and on-farm variable costs. Further field studies are needed for the validation of these results. [ABSTRACT FROM AUTHOR]

Highlights - Efficacy of UAVs and emergence predictive models for weed control has been confirmed. - Combination of time-specific and site-specific weed control provides optimal results. - Use of timely prescription maps can substantially reduce herbicide use. Remote sensing using unmanned aerial vehicles (UAVs) for weed detection is a valuable asset in agriculture and is vastly used for site-specific weed control. Alongside site-specific methods, time-specific weed control is another critical aspect of precision weed control where, by using different models, it is possible to determine the time of weed species emergence. In this study, site-specific and time-specific weed control methods were combined to explore their collective benefits for precision weed control. Using the AlertInf model, which is a weed emergence prediction model, the cumulative emergence of Sorghum halepense was calculated, following the selection of the best date for UAV survey when the emergence was predicted to be at 96%. The survey was executed using a UAV with visible range sensors, resulting in an orthophoto with a resolution of 3 cm, allowing for good weed detection. The orthophoto was post-processed using two separate methods: an artificial neural network (ANN) and the visible atmospherically resistant index (VARI) to discriminate between the weeds, the crop and the soil. Finally, a model was applied for the creation of prescription maps with different cell sizes (0.25 m2, 2 m2, and 3 m2) and with three different decision-making thresholds based on pixels identified as weeds (>1%, >5%, and >10%). Additionally, the potential savings in herbicide use were assessed using two herbicides (Equip and Titus Mais Extra) as examples. The results show that both classification methods have a high overall accuracy of 98.6% for ANN and 98.1% for VARI, with the ANN having much better results concerning user/producer accuracy and Cohen's Kappa value (k=83.7 ANN and k=72 VARI). The reduction percentage of the area to be sprayed ranged from 65.29% to 93.35% using VARI and from 42.43% to 87.82% using ANN. The potential reduction in herbicide use was found to be dependent on the area. For the Equip herbicide, this reduction ranged from 1.32 L/ha to 0.28 L/ha for the ANN; with VARI the reduction in the amounts used ranged from 0.80 L/ha to 0.15 L/ha. Meanwhile, for Titus Mais Extra herbicide, the reduction ranged from 46.06 g/ha to 8.19 g/ha in amounts used with the ANN; with VARI the reduction in amounts used ranged from 27.77 g/ha to 5.32 g/ha. These preliminary results indicate that combining site-specific and time-specific weed control, has the potential to obtain a significant reduction in herbicide use with direct benefits for the environment and on-farm variable costs. Further field studies are needed for the validation of these results.

Broadcast spraying method is generally used for weed control in agriculture. Therefore, excessive amounts of pesticides are sprayed. In this study, a mobile robot was developed and tested on artificial weed targets for a micro-dose spraying system that works only with weed targets in order to reduce the use of spraying liquid in weed control. A prototype mobile robot consisted of a robotic platform, machine vision and steerable spraying unit was built and controlled by using LabVIEW, and tested to evaluate the feasibility of the spraying system. Greenness method and segmentation algorithm were utilized to extract artificial weed from the background. The artificial weed samples were treated according to their coordinates by the servo-based micro-dose spraying needle nozzle. The experiments were conducted at the speeds of 0.42, 0.54, 0.66, 0.78 and 0.90 km h-1 to evaluate the performance of the spraying system under laboratory conditions. The tracking and targeting performances of the mobile spraying system were observed visually. Consumption, the amount of deposition and coverage rate experiments were conducted by using graduated cups, filter papers and water-sensitive papers to evaluate the spraying efficiency of the system at 200 kPa spraying pressure. As a result, the targeted micro-dose spraying application saved about 95% application volume comparing with the broadcast spraying method. Higher spraying efficiency was determined at the middle locations than the edge locations according to the amount of deposition and coverage rate results. The developed servo-based target oriented weed control system was tested experimentally and found to be very efficient.

Copyright of Gesunde Pflanzen is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

This review is designed to address various alternative weed-control practices and their possibilities in the fruit orchard in terms of sustainability. Correct weed management and maintenance of adequate orchard biodiversity are crucial for sustainable orchard soil management. The key is to practice an alternative weed-management approach (single or integrated) rather than to use possibly harmful chemicals only. Integration of modern equipment with a shallow tillage system can provide effective weed control in tree rows, including optimised tree performance and soil biodiversity. Living mulch suppresses weeds and enhances orchard biodiversity, while selection of less competitive and less pest-attracting species is crucial. Plastic covers offer long-term weed control, but additional nutrient amendments are required to maintain the balanced fertility of the soil. Wood chip mulch is suggested where the materials are available on or near the farm, and where there is lower incidence of perennial weeds. High pressure water and robotic systems are still in their infancy for fruit orchards, and required more research to confirm their efficiency. [ABSTRACT FROM AUTHOR]

Copyright of Julius-Kühn-Archiv is the property of Julius Kuehn Institut and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Precision application of specific herbicides to susceptible weeds can significantly save herbicide. This is the first study evaluating the performances of precision sprayer for weed control in turf based on the herbicide weed control spectrum in field conditions. The results showed that EfficientNet-v2 and ResNet never fall below 0.992 for discriminating and detecting the grid cells encompassing weeds susceptible to ACCase-inhibiting and synthetic auxin herbicides. MCPA, a synthetic auxin herbicide, is used to evaluate the performance of the developed smart sprayer for precision control of broadleaf weeds in dormant bermudagrass turf. The developed smart sprayer prototype detected and sprayed every grid cell containing broadleaf weeds in field experiments. Compared to the broadcast application, precision spraying of MCPA provided the same level of control of broadleaf weeds. By 18 days after treatment (DAT), the nontreated control had 13 weeds no. m−2, while the plots that received broadcast and precision spraying had 0 and 1 broadleaf weed plant no. m−2, respectively. Precision herbicide application according to the herbicide weed control spectrum (HWCS) with the developed smart sprayer provided the same level of broadleaf weed control and could save more herbicides compared to an approach without discriminating weed species. Overall, these findings clearly indicated that the developed smart sprayer prototype could effectively detect, discriminate, and spray herbicides onto the grid cells containing target weeds based on the HWCS. • Discriminating weed species based on their susceptibility to herbicides allows targeted herbicide application. • A smart sprayer prototype was designed and developed for precision herbicide application on turf. • Precision herbicide application with the smart sprayer achieved weed control equivalent to broadcast application. [ABSTRACT FROM AUTHOR]

Harrowing is often used to reduce weed competition, generally using a constant intensity across a whole field. The efficacy of weed harrowing in wheat and barley can be optimized, if site-specific conditions of soil, weed infestation and crop growth stage are taken into account. This study aimed to develop and test an algorithm to automatically adjust the harrowing intensity by varying the tine angle and number of passes. The field variability of crop leaf cover, weed density and soil density was acquired with geo-referenced sensors to investigate the harrowing selectivity and crop recovery. Crop leaf cover and weed density were assessed using bispectral cameras through differential images analysis. The draught force of the soil opposite to the direction of travel was measured with electronic load cell sensor connected to a rigid tine mounted in front of the harrow. Optimal harrowing intensity levels were derived in previously implemented experiments, based on the weed control efficacy and yield gain. The assessments of crop leaf cover, weed density and soil density were combined via rules with the aforementioned optimal intensities, in a linguistic fuzzy inference system (LFIS). The system was evaluated in two field experiments that compared constant intensities with variable intensities inferred by the system. A higher weed density reduction could be achieved when the harrowing intensity was not kept constant along the cultivated plot. Varying the intensity tended to reduce the crop leaf cover, though slightly improving crop yield. A real-time intensity adjustment with this system is achievable, if the cameras are attached in the front and at the rear or sides of the harrow.

Harrowing is often used to reduce weed competition, generally using a constant intensity across a whole field. The efficacy of weed harrowing in wheat and barley can be optimized, if site-specific conditions of soil, weed infestation and crop growth stage are taken into account. This study aimed to develop and test an algorithm to automatically adjust the harrowing intensity by varying the tine angle and number of passes. The field variability of crop leaf cover, weed density and soil density was acquired with geo-referenced sensors to investigate the harrowing selectivity and crop recovery. Crop leaf cover and weed density were assessed using bispectral cameras through differential images analysis. The draught force of the soil opposite to the direction of travel was measured with electronic load cell sensor connected to a rigid tine mounted in front of the harrow. Optimal harrowing intensity levels were derived in previously implemented experiments, based on the weed control efficacy and yield gain. The assessments of crop leaf cover, weed density and soil density were combined via rules with the aforementioned optimal intensities, in a linguistic fuzzy inference system (LFIS). The system was evaluated in two field experiments that compared constant intensities with variable intensities inferred by the system. A higher weed density reduction could be achieved when the harrowing intensity was not kept constant along the cultivated plot. Varying the intensity tended to reduce the crop leaf cover, though slightly improving crop yield. A real-time intensity adjustment with this system is achievable, if the cameras are attached in the front and at the rear or sides of the harrow. [ABSTRACT FROM AUTHOR]