RESUMO
The knowledge that precision weed control in agricultural fields can reduce waste and increase productivity has led to research into autonomous machines capable of detecting and removing weeds in real time. One of the driving factors for weed detection is to develop alternatives to herbicides, which are becoming less effective as weed species develop resistance. Advances in deep learning technology have significantly improved the robustness of weed detection tasks. However, deep learning algorithms often require extensive computational resources, typically found in powerful computers that are not suitable for deployment in robotic platforms. Most ground rovers and UAVs utilize embedded computers that are portable but limited in performance. This necessitates research into deep learning models that are computationally lightweight enough to function in embedded computers for real-time applications while still maintaining a base level of detection accuracy. This paper evaluates the weed detection performance of three real-time-capable deep learning models, YOLOv4, EfficientDet, and CenterNet, when run on a deep-learning-enabled embedded computer, an Nvidia Jetson Xavier AGX. We tested the accuracy of the models in detecting 13 different species of weeds and assesses their real-time viability through their inference speeds on an embedded computer compared to a powerful deep learning PC. The results showed that YOLOv4 performed better than the other models, achieving an average inference speed of 80 ms per image and 14 frames per second on a video when run on an imbedded computer, while maintaining a mean average precision of 93.4% at a 50% IoU threshold. Furthermore, recognizing that some real-world applications may require even greater speed, and that the detection program would not be the only task running on the embedded computer, a lightweight version of the YOLOv4 model, YOLOv4-tiny, was tested for improved performance in an embedded computer. YOLOv4-tiny impressively achieved an average inference speed of 24.5 ms per image and 52 frames per second, albeit with a slightly reduced mean average precision of 89% at a 50% IoU threshold, making it an ideal choice for real-time weed detection.
RESUMO
Precision weeding can significantly reduce or even eliminate the use of herbicides in farming. To achieve high-precision, individual targeting of weeds, high-speed, low-cost plant identification is essential. Our system using the red, green, and near-infrared reflectance, combined with a size differentiation method, is used to identify crops and weeds in lettuce fields. Illumination is provided by LED arrays at 525, 650, and 850 nm, and images are captured in a single-shot using a modified RGB camera. A kinematic stereo method is utilised to compensate for parallax error in images and provide accurate location data of plants. The system was verified in field trials across three lettuce fields at varying growth stages from 0.5 to 10 km/h. In-field results showed weed and crop identification rates of 56% and 69%, respectively. Post-trial processing resulted in average weed and crop identifications of 81% and 88%, respectively.
RESUMO
To avoid excessive use of herbicides in the weeding operations of Peucedani Radix, a common Chinese herb, a precision seedling avoidance and weeding agricultural robot was designed for the targeted spraying of herbicides. The robot uses YOLOV5 combined with ExG feature segmentation to detect Peucedani Radix and weeds and obtain their corresponding morphological centers. Optimal seedling avoidance and precise herbicide spraying trajectories are generated using a PSO-Bezier algorithm based on the morphological characteristics of Peucedani Radix. Seedling avoidance trajectories and spraying operations are executed using a parallel manipulator with spraying devices. The validation experiments showed that the precision and recall of Peucedani Radix detection were 98.7% and 88.2%, respectively, and the weed segmentation rate could reach 95% when the minimum connected domain was 50. In the actual Peucedani Radix field spraying operation, the success rate of field precision seedling avoidance herbicide spraying was 80.5%, the collision rate between the end actuator of the parallel manipulator and Peucedani Radix was 4%, and the average running time of the parallel manipulator for precision herbicide spraying on a single weed was 2 s. This study can enrich the theoretical basis of targeted weed control and provide reference for similar studies.
RESUMO
Due to the different weed characteristics in peanut fields at different weeding periods, there is an urgent need to study a general model of peanut and weed detection and identification applicable to different weeding periods in order to adapt to the development of mechanical intelligent weeding in fields. To this end, we propose a BEM-YOLOv7-tiny target detection model for peanuts and weeds identification and localization at different weeding periods to achieve mechanical intelligent weeding in peanut fields at different weeding periods. The ECA and MHSA modules were used to enhance the extraction of target features and the focus on predicted targets, respectively, the BiFPN module was used to enhance the feature transfer between network layers, and the SIoU loss function was used to increase the convergence speed and efficiency of model training and to improve the detection performance of the model in the field. The experimental results showed that the precision, recall, mAP and F1 values of the BEM-YOLOv7-tiny model were improved by 1.6%, 4.9%, 4.4% and 3.2% for weed targets and 1.0%, 2.4%, 2.2% and 1.7% for all targets compared with the original YOLOv7-tiny. The experimental results of positioning error show that the peanut positioning offset error detected by BEM-YOLOv7-tiny is less than 16 pixels, and the detection speed is 33.8 f/s, which meets the requirements of real-time seedling grass detection and positioning in the field. It provides preliminary technical support for intelligent mechanical weeding in peanut fields at different stages.
RESUMO
Deep learning techniques have made great progress in the field of target detection in recent years, making it possible to accurately identify plants in complex environments in agricultural fields. This project combines deep learning algorithms with spraying technology to design a machine vision precision real-time targeting spraying system for field scenarios. Firstly, the overall structure scheme of the system consisting of image acquisition and recognition module, electronically controlled spray module and pressure-stabilized pesticide supply module was proposed. After that, based on the target detection model YOLOv5s, the model is lightened and improved by replacing the backbone network and adding an attention mechanism. Based on this, a grille decision control algorithm for solenoid valve group on-off was designed, while common malignant weeds were selected as objects to produce data sets and complete model training. Finally, the deployment of the hardware system and detection model on the electric spray bar sprayer was completed, and field trials were conducted at different speeds. The experimental results show that the improved algorithm reduces the model size to 53.57% of the original model with less impact on mAP accuracy, improves FPS by 18.16%. The accuracy of on-target spraying at 2km/h, 3km/h and 4km/h speeds were 90.80%, 86.20% and 79.61%, respectively, and the spraying hit rate decreased as the operating speed increased. Among the hit rate components, the effective recognition rate was significantly affected by speed, while the relative recognition hit rate was less affected.