Advances in ground robotic technologies for site-specific weed management in precision agriculture: A review

Computers and Electronics in Agriculture, Год журнала: 2024, Номер 225, С. 109363 - 109363

Опубликована: Авг. 22, 2024

Язык: Английский

---

Laser weeding: opportunities and challenges for couch grass (Elymus repens (L.) Gould) control

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Май 15, 2024

Abstract Laser weeding may contribute to less dependency on herbicides and soil tillage. Several research commercial projects are underway develop robots equipped with lasers control weeds. Artificial intelligence can be used locate identify weed plants, mirrors direct a laser beam towards the target kill it heat. Unlike chemical mechanical control, only exposes tiny part of field for treatment. leaves behind ashes from burned plants does not disturb soil. Therefore, is an eco-friendly method seedlings. However, perennial weeds regrow belowground parts after destroys aerial shoots. Depletion resources might possible if continuously kills new shoots, but require many treatments. We studied how could destroy widespread aggressive Elymus repens rhizomes were cut into fragments. Plants killed even small dosages energy stopped regrowing. Generally, highest efficacy was achieved when treated at 3-leaf stage.

Язык: Английский

---

Environmental performance of an autonomous laser weeding robot—a case study

The International Journal of Life Cycle Assessment, Год журнала: 2024, Номер 29(6), С. 1021 - 1052

Опубликована: Апрель 8, 2024

Abstract Purpose Challenges in sustainable development envisioned the European Union for agricultural sector require innovation to raise efficiency of production and safety farming processes farmers ensure food consumers. One key productivity factors plant is effective weeding. The WeLASER project aimed develop a high-power autonomous vehicle with lasers control weeds. To be sustainable, invention should have high environmental performance whole life cycle perspective, including its production, use agriculture, end-of-life phase. In publication, assessment (LCA) weeding robot presented. aim was identify weak strong aspects terms provide suggestions improvement. Methods machinery characterized based on technical data provided by developers, relevant literature, Ecoinvent 3.8 database, own calculations. quantitative impacts performed using Simapro tool. For interpretation Recipe 2016 method (egalitarian perspective) applied. Results results show that energy issue related laser-based machine operations most challenging. It climate change indicators fossil fuel depletion. Production phase human toxicity extensive application electronic electric components robot. Conclusion comparison other techniques, weeds shows potential environmentally efficient practice. Achieving perspective requires improvements design, operational features, smart practice enhanced through expertise, guidance, advice.

Язык: Английский

---

Design of and Experimentation on an Intelligent Intra-Row Obstacle Avoidance and Weeding Machine for Orchards

Agriculture, Год журнала: 2025, Номер 15(9), С. 947 - 947

Опубликована: Апрель 27, 2025

Based on the current issues of difficulty in clearing intra-row weeds orchards, inaccurate sensor detection, and inability to adjust row spacing depth, this study designs an intelligent obstacle avoidance weeding machine for orchards. We designed machine’s device, depth-limiting adjustment mechanism, joystick-based shovel, hydraulic system. The device integrates non-contact sensors a mechanical tactile structure, which overcomes instability detection avoids risk collision by parts. shovel can be adapted environments orchards with small plant spacing. combination mechanism realizes flexible avoidance. used Ansys Workbench conduct static vibration modal analyses chassis in-field machine. On basis, through topology optimization, quality is reduced 8%, goal light weight ensures stable operation machinery. To further optimize parameters, we employed Box–Behnken design response surface analysis, coverage as optimization target. systematically explored effects forward speed, cylinder extension retraction speed efficiency. optimal operational parameter determined follows: 0.5 m/s, 11.5 cm/s, 8 cm/s. theoretical analysis scenario simulations, validated performance field experiments. results show that machine, while exhibiting excellent performance, achieve maximum 84.6%. This provides foundation technical support development weeding, great significance achieving orchard management improving fruit yield quality.

Язык: Английский

---

Algorithm for Locating Apical Meristematic Tissue of Weeds Based on YOLO Instance Segmentation

Agronomy, Год журнала: 2024, Номер 14(9), С. 2121 - 2121

Опубликована: Сен. 18, 2024

Laser technology can be used to control weeds by irradiating the apical meristematic tissue (AMT) of when they are still seedlings. Two factors necessary for successful large-scale implementation this technique: ability accurately identify and effectiveness localization algorithm in process. Based on this, study proposes a lightweight weed AMT based YOLO (look only once) instance segmentation. The YOLOv8n-seg network undergoes design enhancement integrating FasterNet as its backbone, resulting F-YOLOv8n-seg model. This modification effectively reduces number parameters computational demands during convolution process, thereby achieving more efficient Subsequently, is combined with connected domain analysis (CDA), yielding F-YOLOv8n-seg-CDA integration enables precise calculating center-of-mass coordinates domains. experimental results indicate that optimized model significantly outperforms original model; floating-point computations 26.7% size 38.2%. In particular, calculation decreased 8.9 GFLOPs, lowered 4.2 MB. Comparing improved against YOLOv5s-seg YOLOv10n-seg, it lighter. Furthermore, exhibits exceptional segmentation accuracy, 97.2% accuracy rate. Experimental tests conducted five different species demonstrated strong generalization capabilities. detecting these was 81%. Notably, dicotyledonous were detected up 94%. Additionally, achieved an average inference speed 82.9 frames per second. These suitable real-time detection tissues across multiple species. impact distinctive variations morphology identifying location weeds. It discovered monocotyledonous differed terms effect, having higher than discovery offer novel insights avenues future investigation into identification

Язык: Английский

---

A novel mechanical-laser collaborative intra-row weeding prototype: structural design and optimization, weeding knife simulation and laser weeding experiment

Frontiers in Plant Science, Год журнала: 2024, Номер 15

Опубликована: Окт. 30, 2024

The competition between intra-row weeds and cultivated vegetables for nutrients is a major contributor crop yield reduction. Compared with manual weeding, intelligent robots can improve the efficiency of weeding operations.

Язык: Английский

---