Weed Control Costs Research Articles

Exploring the semi-supervised learning for weed detection in wheat

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.

Non-chemical weed management: Harnessing flame weeding for effective weed control

The goal of the current study was to create and assess the effectiveness of a hand-pulled ergonomically designed flame weeder. The developed weeder was tested in the field at three operating pressures (20, 30 and 40 Psi) and forward speeds (1.00, 1.25 and 1.50 km/h) to study their effects on plant damage, survival rates, weight preservation rates, weed management effectiveness, soil temperatures, and gas and energy consumption. Thereafter, at optimized values of forward speed and operating pressure, a comparative assessment of flame weeding with traditional methods (mechanical and manual weeding) was done in terms of weed control effectiveness, operational time, energy consumption, and cost of operation. Results showed that the optimal performance of the designed flame weeder was achieved when operated at a speed of 1 km/h and an operating pressure of 40 psi. The survival rate, weight preservation rate, weed control efficiency, change in soil temperature, recovery rate, plant damage, gas consumption, and energy consumption were observed to be 27.3 %, 32.5 %, 91.1 %, 40.74 °C, 8.5 %, 2.2 %, 4.05 kg/h, and 2500.24 MJ/ha, respectively, at optimized values of forward speed (1.00 km/h) and operating pressure (40 Psi). The actual field capacity, field efficiency and operating cost of the flame weeder were 0.0755 ha/h, 94.94 %, and 3620.81 ₹/ha, respectively. Hand weeding had the best level of weed control effectiveness, but it was a laborious, time-consuming process. When compared to manual weeding, flame weeding was 50.42 % cheaper and 94.82 % faster.

Development and Validation of Avena Integrated Management (AIM): A Bioeconomic Decision Support Tool for Wild Oat Management in Australian Grain Production Systems

Abstract Wild oat is a long-standing weed problem in Australian grain cropping systems, potentially reducing the yield and quality of winter grain crops significantly. The effective management of wild oat requires an integrated approach of diverse control techniques that suit specific crops and cropping situations. This research aimed to construct and validate a bioeconomic model that enables the simulation and integration of weed control technologies for wild oat in grain production systems. The Avena spp. integrated management (AIM) model was developed with a simple interface to provide outputs of biological and economic data (crop yields, weed control costs, emerged weeds, weed seedbank, gross margins) on wild oat management data in a cropping rotation. Uniquely, the AIM was validated against real-world data on wild oat management in a wheat and sorghum cropping rotation, where the model was able to reproduce the patterns of wild oat population changes as influenced by weed control and agronomic practices. Correlation coefficients for 12 comparison scenarios ranged between 0.55 and 0.96. With accurate parameterization, AIM is thus able to make useful predictions on the effectiveness of individual and integrated weed management tactics for wild oat control in grain cropping systems.

Weed Recognition at Soybean Seedling Stage Based on YOLOV8nGP + NExG Algorithm

The high cost of manual weed control and the overuse of herbicides restrict the yield and quality of soybean. Intelligent mechanical weeding and precise application of pesticides can be used as effective alternatives for weed control in the field, and these require accurate distinction between crops and weeds. In this paper, images of soybean seedlings and weeds in different growth areas are used as datasets. In the aspect of soybean recognition, this paper designs a YOLOv8nGP algorithm with a backbone network optimisation based on GhostNet and an unconstrained pruning method with a 60% pruning rate. Compared with the original YOLOv8n, the YOLOv8nGP improves the Precision (P), Recall (R), and F1 metrics by 1.1% each, reduces the model size by 3.6 mb, and the inference time was 2.2 ms, which could meet the real-time requirements of field operations. In terms of weed recognition, this study utilises an image segmentation method based on the Normalized Excess Green Index (NExG). After filtering the soybean seedlings, the green parts of the image are extracted for weed recognition, which reduces the dependence on the diversity of the weed datasets. This study combines deep learning with traditional algorithms, which provides a new solution for weed recognition of soybean seedlings.

Evaluation of two deep learning-based approaches for detecting weeds growing in cabbage.

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. 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. 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.

Sustainability evaluation of contrasting milpa systems in the Yucatán Peninsula, Mexico

AbstractThe milpa agroecosystem is an intercropping of maize, beans, squash and other crops, developed in Mesoamerica, and its adoption is widely variable across climates and regions. An example of particular interest is the Yucatan Peninsula in Mexico, which holds highly diverse milpas, drawing on ancestral Mayan knowledge. Traditional milpas have been described as sustainable resource management models, based on long rotations within a slash-and-burn cycle in forest areas. Nevertheless, due to modernization and intensification processes, new variants of the approach have appeared. The objective of this study was to evaluate the sustainability of three milpa systems (traditional, continuous, and mechanized) in four case studies across the Peninsula, with emphasis on food self-sufficiency, social inclusion and adoption of innovations promoted by a development project. The Framework for the Evaluation of Agroecosystems using Indicators (MESMIS, for its Spanish acronym) was used for its flexible, participatory approach. A common group of indicators was developed despite regional differences between study cases, with a high level of farmer participation throughout the iterative process. The results show lower crop yields in traditional systems, but with lower inputs costs and pesticide use. In contrast, continuous milpas had higher value in terms of crop diversity, food security, social inclusion, and innovation adoption. Mechanized milpas had lower weed control costs. Profitability of cash crops and the proportion of forest were high in all systems. Highly adopted innovations across milpa types and study cases included spatial crop arrangement and the use of residues as mulches. However, most innovations are not adapted to local conditions, and do not address climate change. Further, women and youth participation is low, especially in traditional systems.

Investigating image-based fallow weed detection performance on Raphanus sativus and Avena sativa at speeds up to 30 km h−1

Site-specific weed control (SSWC) can provide considerable reductions in weed control costs and herbicide usage. Despite the promise of machine vision for SSWC systems and the importance of ground speed in weed control efficacy, there has been little investigation of the role of ground speed and camera characteristics on weed detection performance. Here, we compare the performance of four camera/software combinations using the open-source OpenWeedLocator platform – (1) default settings on a Raspberry Pi HQ camera (HQ1); (2) optimised software settings on a HQ camera (HQ2), (3) optimised software settings on the Raspberry Pi v2 camera (V2); and (4) a global shutter Arducam AR0234 camera (ARD) – at speeds ranging from 5 km h−1 to 30 km h−1. A combined excess green (ExG) and hue, saturation, value (HSV) thresholding algorithm was used for testing under fallow conditions using tillage radish (Raphanus sativus) and forage oats (Avena sativa) as representative broadleaf and grass weeds, respectively. ARD demonstrated the highest recall among camera systems, with up to 100 % recall for tillage radish and 91.6 % for forage oats at 5 km h−1. At 30 km h−1 ARD recall for forage oats declined (P < 0.05) by 12.4 %. There was no significant change for tillage radish. All cameras experienced a decrease in recall as speed increased. The highest rate of decrease was observed for HQ1 with 1.12 % and 0.90 % reductions in recall for every km/h increase in speed for tillage radish and forage oats, respectively. Detection of the ‘grassy’ forage oats was worse (P < 0.05) than the broadleaved tillage radish for all cameras. Despite the variations in recall, HQ1, HQ2, and V2 maintained near-perfect precision at all tested speeds. The variable effect of ground speed and camera system on detection performance of grass and broadleaf weeds, indicates that careful hardware and software considerations must be made when developing SSWC systems.

Rationalizing Herbicide Use in Maize within the Framework of Climatic Change and Extreme Hydrometeorological Phenomena

The aim of this study was to investigate specific management methods for controlling weeds under different climate conditions by reducing the risk of herbicide resistance in maize. Herbicide trials were placed at two locations during 2017 and 2018 and organized in a randomized block design with four replicates of all herbicide combinations (terbuthylazine—pre-emergence and mesotrione post-emergence treatment, terbuthylazine + mesotrione in post-emergence treatment, and terbuthylazine—pre-emergence and dicamba in post-emergence treatment). In the trials, 13 specific weeds were observed. High effectiveness was achieved when using commercial preparations with two active ingredients (terbuthylazine and mesotrione) in both years of research. As the year 2018 featured a higher amount of precipitation, uneven weed sprouting occurred, which resulted in a large number of Setaria glauca (18.50 No m−2), Solanum nigrum (16.50 No m−2), Datura stramonium (13.75 No m−2), and Chenopodium hybridum (10.50 No m−2) plants. Herbicide phytotoxicity was not expressed in maize. Ambrosia artemisiifolia, Solanum nigrum, and Datura stramonium had the highest competitive index (CI 10). The highest maize yield was observed under the application of terbuthylazine + mesotrione in post-emergence (14.223 t/ha). This combination led to the highest weed control costs (35.60 EUR/ha) in 2018, while in 2017, the yield was 12.829 t/ha, with a control cost of 33.99 EUR/ha.

Soybean-rice rotation in southern Brazil: economics advantages

The aim of this paper is to show a step-by-step method and empirical data for evaluating the economic advantages of introducing soybean in rotation with irrigated rice in the Brazilian southern lowlands by comparing two groups of farms: (01) farms with rice’s monoculture; (02) farms with soybean and rice rotation in the last 5 years. As a result, we found that the introduction of soybeans in the irrigated rice production system reduces the costs of soil preparation (41.45%), weed control costs (49.60%) and exposure to risk (27,61%). Furthermore, through rotation farms might improve the results of commercialization by 0,66% in comparison with monocultural farms. On the other hand, the rotation system increases fertilization costs (13.34%), reduces the gains with cattle grazing in the fallow period (0,97%) and the gross revenue per unit of area (5.62%). In sum the rotation is a viable production strategy, generating an increase of 9,68% in the financial return as compared to the traditional rice’s monocultural agricultural system.

Metabolic Energy Consumption during Green Area Management

The energy consumption measurement is important to carry out a correct risk evaluation of workers during green area management with the scope to achieve workstyle improvements. In contrast to sporting activities, few studies have been conducted on the assessment of physical fatigue by determining the functional parameters of the human body and oxygen consumption in this sector. This study aims to measure the energy cost of weed control using a wearable telemetry system Cosmed K4. For this purpose, twelve male workers, grouped into three groups of four workers each, were monitored during the work of weed control carried out with a brush cutter (by testing three different cutting heads). The monitoring period lasted 18 min including a 5-min rest period at the end of work. This study shows how the use of facilitating tools such as brush cutters contributes to getting low energy metabolism rate values, in tests performed equally on average to 119 Wm−2.

Weed Seed Banks in Intensive Farmland and the Influence of Tillage, Field Position, and Sown Flower Strips

Agricultural intensification has caused once diverse arable fields to become species-poor. Their seed banks, which are fundamental for re-establishment and maintenance of plant communities in such repeatedly disturbed environments, are now largely depleted. In order to advise farmers on the successful implementation of agri-environmental measures, as well as reduce potential subsequent costs of continued weed control, understanding seed bank dynamics in relation to aboveground vegetation is essential. We (1) investigated the change in seed bank composition in the field edge and the interior, and (2) analyzed the seed bank in flower strips and adjacent fields in relation to the aboveground vegetation on intensively managed arable farms across Germany. Low-tillage systems contained more plant species and higher seed densities in the seed bank than regularly ploughed fields. Species diversity at the field edge was higher than in the field interior, with a continuous decrease in the number of species and seed density within the first 2 m from the edge. Flower strips can lead to an enrichment of the seed bank, but it is driven by the strong rise in a few common species such as Chenopodium album. To cultivate successful flower strips, we recommend close onsite monitoring, as well as rapid intervention in the case of weed infestation.

BASIC CONSTRUCTIONS OF SECTIONS FOR FERTILIZER APPLICATION IN SECTION OF STRIP-TILL TECHNOLOGY

Strip-till technology was formed in the USA on a short corn-soybean crop rotation. Both crops are harvested in September-October and thereby exclude the appearance of post-harvest weeds. In addition, both crops are grown on wide-row crops, and this prompted the emergence of strip cultivation technology. Strip-till technology is an alternative tillage option between traditional plowing and direct seeding. When applying this technology, the soil is deeply loosened only within the row, the rest of the field remains untreated. Structurally, each section for strip soil cultivation consists of such working bodies as a frame, a front slotted disk that cuts the soil surface and plant remains, and also withstands the depth of cultivation. Cutting discs on one rack clear the row of plant remains and lightly mulch the soil surface. The section rack performs tillage to a depth of up to 30 cm, and if additional equipment is available, it can provide fertilizers, including at two different depth values, and the rear roller compacts and grinds the soil surface. The use of this technology makes it possible to reduce the number of operations in one pass of the unit. Also, the above-mentioned tillage technology can solve such problems as high weediness, low fertility of soils with insufficient content of organic substances, the presence of a short growing season, and can also be used in arid regions and soils prone to wind erosion. The disadvantage of this technology is the need to apply a continuous action herbicide. But these economic costs of weed control are not significant in the total cost of tillage. It should be noted that herbicides are also applied according to traditional technology.

English

As Ghana prepares to commence commercialization of Genetically Modified (GM) crops, this paper reports findings of a survey that sought the views of smallholder farmers in northern Ghana on the likely prospects and constraints of GM crops cultivation. Multi-stage sampling techniques were employed to sample 360 smallholder farmers across ten districts in northern Ghana with interviews, guided by semi-structured questionnaires, and focus group discussions used to gather data for the study. Descriptive statistics and Kendall&rsquo;s coefficient of concordance were applied to analyze the data gathered. The prospects of GM technology according to the smallholder farmers surveyed were that GM technology can be used to breed drought-tolerant, early maturing, and high yielding local crop varieties and also help reduce the cost of weed, pest, and disease control. While their perceived constraints to the cultivation of GM crops were the probably high cost of GM seed, unreliable supply of GM seed, likely failure of regulatory agencies, and possible environmental and health risks among others. It is recommended to the Ministry of Food and Agriculture (MOFA) and other relevant organizations to address farmers&rsquo; concerns and reservations over the possible high cost of GM seeds and the fear that GM seed supply might not be reliable. Key words: Prospects, constraints, GM crops, agricultural biotechnology, smallholder farmers, northern region of Ghana, biosafety.

Advances in site‐specific weed management in agriculture—A review

AbstractThe developments of information and automation technologies have opened a new era for weed management to fit physical and chemical control treatments to the spatial and temporal heterogeneity of weed distributions in agricultural fields. This review describes the technologies of site‐specific weed management (SSWM) systems, evaluates their ecological and economic benefits and gives a perspective for the implementation in practical farming. Sensor technologies including 3D cameras, multispectral imaging and Artificial Intelligence (AI) for weed classification and computer‐based decision algorithms are described in combination with precise spraying and hoeing operations. Those treatments are targeted for patches of weeds or individual weed plants. Cameras can also guide inter‐row hoes precisely in the centre between two crop rows at much higher driving speed. Camera‐guided hoeing increased selectivity and weed control efficacy compared with manual steered hoeing. Robots combine those guiding systems with in‐row hoeing or spot spraying systems that can selectively control individual weeds within crop rows. Results with patch spraying show at least 50% saving of herbicides in various crops without causing additional costs for weed control in the following years. A challenge with these technologies is the interoperability of sensing and controllers. Most of the current SSWM technologies use their own IT protocols that do not allow connecting different sensors and implements. Plug &amp; play standards for linking detection, decision making and weeding would improve the adoption of new SSWM technologies and reduce operational costs. An important impact of SSWM is the potential contribution to the EU‐Green Deal targets to reduce pesticide use and increase biodiversity. However, further on‐farm research is needed for integrating those technologies into agricultural practice.

A preliminary study of ‘Tidy Field Technology’ to assess growth, development and weed control in direct-seeded rice

Direct-seeded rice (DSR) is an important method to achieve simplified and mechanized rice planting in economical manner. However, proper weed management is a major limiting factor in the development of DSR. ‘Tidy field technology’ (TFT) is a new non-transgenic herbicide-resistant (HR) cropping technology, which refers to ‘herbicide-resistant crops combined with broad-spectrum herbicide’, and has been widely confirmed to effectively suppress weed populations in agriculture. However, it is rarely applied in DSR production in China. Therefore, field experiments were conducted to determine the effects of TFT on rice growth and weed control in 2019 and 2020. In this study, two rice varieties, Huanghuazhan-HHZ (Vc) and herbicide-resistant HHZ (Vr) were adopted and practiced with broad-spectrum herbicide management (Hb) and conventional herbicide management (Hc). An equal amount of water was sprayed on Vc as a control (CK). Thus, there were five treatments: VrHb, VrHc, VcHb, VcHc and CK. The grain yields of VrHb was increased by 8.4% and 160.3% in 2019, and 16.0% and 226.6% in 2020 compared with VrHc and CK, respectively. Furthermore, there were no significant differences in grain yield between VrHb and VcHc, but the total economic benefit of VrHb was significantly higher than that of VcHc. In addition, the tiller number, leaf area index (LAI), and aboveground total dry weight (TDW) of VrHb were significantly higher than VrHc, VcHc and CK. In contrast, weed species and density under VrHb were significantly lower than those under VrHc, VcHc and CK. Overall, TFT (VrHb) could replace conventional weed management for DSR due to a better weed control effect and higher economic benefit with similar or even higher grain yield. However, it is necessary to explore the effects of TFT residue on soil physical and chemical properties, microorganisms and the growth of subsequent crops in the near future. • Tidy field technology (TFT) was a new non-transgenic herbicide-resistant technology. • Direct-seeded rice (DSR) had the highest yield and net economic benefit under TFT. • High yield of DSR under TFT was due to better weeds control and crop establishment. • High economic benefit of DSR under TFT was due to the reduced cost of weed control.

Development of a Three-Dimensional Plant Localization Technique for Automatic Differentiation of Soybean from Intra-Row Weeds

Soybean is a legume that is grown worldwide for its edible bean. Intra-row weeds greatly hinder the normal growth of soybeans. The continuous emergence of herbicide-resistant weeds and the increasing labor costs of weed control are affecting the profitability of growers. The existing cultivation technology cannot control the weeds in the crop row which are highly competitive with the soybean in early growth stages. There is an urgent need to develop an automated weeding technology for intra-row weed control. The prerequisite for performing weeding operations is to accurately determine the plant location in the field. The purpose of this study is to develop a plant localization technique based on systemic crop signalling to automatically detect the appearance of soybean. Rhodamine B (Rh-B) is a signalling compound with a unique fluorescent appearance. Different concentrations of Rh-B were applied to soybean based on seed treatment for various durations prior to planting. The potential impact of Rh-B on seedling growth in the outdoor environment was evaluated. Both 60 and 120 ppm of Rh-B were safe for soybean plants. Higher doses of Rh-B resulted in greater absorption. A three-dimensional plant localization algorithm was developed by analyzing the fluorescence images of multiple views of plants. The soybean location was successfully determined with the accuracy of 97%. The Rh-B in soybean plants successfully created a machine-sensible signal that can be used to enhance weed/crop differentiation, which is helpful for performing automatic weeding tasks in weeders.

SURVEY, IDENTIFICATION AND CONTROL OF WEEDS ASSOCIATED WITH MULBERRY (Morus spp.) IN THE PROVINCES OF AURORA, LA UNION, ILOCOS SUR AND ILOCOS NORTE, PHILIPPINES

Surveys, collections and identification of weeds were done in the sericulture projects of Sta. Maria, Ilocos Sur; Marcos, Ilocos Norte; Aurora, Aurora; and aat the SRDI, Bacnotan, La Union, in both wet and dry seasons. A total of 75 species of weeds were collected and identified. Survey, collection and identification of weeds in the sericulture projects of Sta Maria, Ilocos Sur;Marcos, Ilocos Norte; Ma. Aurora, Aurora; and at the Sericulture Research Development Institute in Bacnotan, La Union were done through quadrant sampling. Dominant weeds during the wet season in Sta Maria, Aurora, and Bacnotan were Digitaria sp., Fimbristylis littoralis, Ageratum conyzoides and Cyperus rotundus while during dry season, Diqitaria sp., Chloris sp., Centrosema pubescens and Dactyloctenium aeqyptium were the dominant weed species, respectively.&#x0D; On a per hectare basis, results showed mulching was found to be the most economical and beneficial weed control method. Total cost of weed control was P5,080.00 as against herbicide application with P7,224.00 and hand-weeding at P7,620.00.

Challenging sheep grazing in orchards: Changes in nutrition, performance, and the health of animals and the effects on the vegetation and soil

AbstractWe proposed a mixture of animal husbandry and fruit tree industry: livestock grazing under an orchard layer to increase agricultural productivity per area and reduce the costs of feed, fertilizer, and weed control. Lambs were reared in‐house for 1 month and then moved to a pear orchard for 1 month of grazing. We recorded the changes in body weight, blood properties (e.g., albumin, cholinesterase, and nonesterified fatty acids), and parasitic infection to compare the in‐housed lambs with the grazing lambs in the orchard. Additionally, we conducted field investigations for changes in vegetation and soil properties before and after grazing. Grazing caused significant increases in body weight, helminth eggs in dung, and soil water retention, and significant decreases in plant biomass, grass digestibility, and soil hardness. In contrast, the nutritional values of the feed, blood properties, and soil chemical properties (e.g., pH, P, N, and C) were not significantly changed after grazing. These results highlighted that the fruit tree industry combined with livestock production would be an acceptable management approach for Japanese agriculture, although further case studies are required.

Image based thermal sensing for glyphosate resistant weed identification in greenhouse conditions

In mid-western states, studies have estimated that herbicide-resistant weed management has increased weed control costs by $20–$40 per acre. From 2019 to 2020, greenhouse research was conducted to investigate the canopy temperature response of waterhemp (Amaranthusrudis), kochia (Kochiascoparia), common ragweed (Ambrosiaartemisiifolia), horseweed (Conyzacanadensis),and Palmer amaranth (Amaranthuspalmeri),after glyphosate application to identify glyphosate resistance. In theseexperiments,thermal images were captured of randomized glyphosate-resistant and -susceptible collections of each weed species. The weed canopies' temperature values were extracted from the thermal images andsubmitted to statistical testing. Various classifiers were established to discriminatebetween resistant and susceptible biotypes.Results showed glyphosate-resistant horseweed was the only biotype under greenhouse conditions reliably classified with the resistance trait, 89% accuracy, using cooler temperature signatures than its susceptible counterpart as the discrimination characteristic. The unstable performance from thermal imagery further suggested canopy temperature data was not a reliable predictor of glyphosate resistance. Investigation of additional multispectral sensing technologies or analysis models should be evaluated for more reliable identification of glyphosate-resistant weed species.

Development of Deep Learning-Based Variable Rate Agrochemical Spraying System for Targeted Weeds Control in Strawberry Crop

Agrochemical application is an important tool in the agricultural industry for the protection of crops. Agrochemical application with conventional sprayers results in the waste of applied agrochemicals, which not only increases financial losses but also contaminates the environment. Targeted agrochemical sprayers using smart control systems can substantially decrease the chemical input, weed control cost, and destructive environmental contamination. A variable rate spraying system was developed using deep learning methods for the development of new models to classify weeds and to accurately spray on desired weeds target. Laboratory and field experiments were conducted to assess the sprayer performance for weed classification and precise spraying of the target weeds using three classification CNNs (Convolutional Neural Networks) models. The DCNNs models (AlexNet, VGG-16, and GoogleNet) were trained using a dataset containing a total of 12,443 images captured from the strawberry field (4200 images with spotted spurge, 4265 images with Shepherd’s purse, and 4178 strawberry plants). The VGG-16 model attained higher values of precision, recall and F1-score as compared to AlexNet and GoogleNet. Additionally VGG-16 model recorded higher percentage of completely sprayed weeds target (CS = 93%) values. Overall in all experiments, VGG-16 performed better than AlexNet and GoogleNet for real-time weeds target classification and precision spraying. The experiments results revealed that the Sprayer performance decreased with the increase of sprayer traveling speed above 3 km/h. Experimental results recommended that the sprayer with the VGG-16 model can achieve high performance that makes it more ideal for a real-time spraying application. It is concluded that the advanced variable rate spraying system has the potential for spot application of agrochemicals to control weeds in a strawberry field. It can reduce the crop input costs and the environmental pollution risks.