Weeds are unwanted and invasive plants that proliferate and compete for resources such as space, water, nutrients, and sunlight, affecting the quality and productivity of the desired crops. Weed detection is crucial for the application of precision agriculture methods and for this purpose machine learning techniques can be used, specifically convolutional neural networks (CNN). This study focuses on the search for CNN architectures used to detect and identify weeds in different crops; 61 articles applying CNN architectures were analyzed during the last five years (2019–2023). The results show the used of different devices to acquire the images for training, such as digital cameras, smartphones, and drone cameras. Additionally, the YOLO family and algorithms are the most widely adopted architectures, followed by VGG, ResNet, Faster R-CNN, AlexNet, and MobileNet, respectively. This study provides an update on CNNs that will serve as a starting point for researchers wishing to implement these weed detection and identification techniques.

Weeds are unwanted and invasive plants characterized by their rapid growth and ability to compete with crops for essential resources such as space, water, nutrients, and sunlight. This competition has a negative impact on crop quality and productivity. To reduce the influence of weeds, precision weeding is used, which uses image sensors and computational algorithms to identify plants and classify weeds using digital images. This study used images of maize (Zea mays L.) to detect four types of weeds (Lolium rigidum, Sonchus oleraceus, Solanum nigrum, and Poa annua). For this purpose, YOLO (You Only Look Once) architectures, YOLOv8s, YOLOv9s, YOLOv10s, and YOLOv11s versions, were trained and compared, along with an architecture based on RT-DETR (Real-Time Detection Transformer), version RT-DETR-1. The YOLO architectures are noted for their real-time detection efficiency, and RT-DETR-l allows evaluation of the impact of an architecture that dispenses with Non-Maximum Suppression (NMS). The YOLOv9s model had the best overall performance, achieving a mAP@0.5 of 0.834 in 60 epochs and an F1-score of 0.78, which demonstrates a optimal balance between accuracy and recall, although with less confidence in its predictions. On the other hand, the RT-DETR-l model stood out for its efficiency in convergence, reaching a competitive performance in only 58 epochs with a mAP@0.5 of 0.828 and an F1-score of 0.80.

Corn (Zea mays L.) is one of the most important cereals worldwide. To maintain crop productivity, it is important to eliminate weeds that compete for nutrients and other resources. The eradication of these causes environmental problems through the use of agrochemicals. The implementation of technology to mitigate this impact is also a challenge. In this work, an artificial vision system was implemented based on the YOLOv5s (You Only Look Once) model, which uses a single convolutional neural network (CNN) that allows differentiating corn from four types of weeds, for which a mobile support structure was built to capture images. The performance of the trained model had a value of mAP@05 (mean Average Precision) at a threshold of 0.5 of 83.6%. A prediction accuracy of 97% and a mAP@05 of 97.5% were obtained for the maize class. For the weed classes, Lolium perenne, Sonchus oleraceus, Solanum nigrum, and Poa annua obtained an accuracy of 86%, 90%, 78%, and 74%, and a mAP@05 of 81.5%, 90.2%, 76.6% and 72.0%, respectively. The results are encouraging for the construction of a precision weeding system.

Weeds are unwanted and invasive plants that proliferate and compete for resources such as space, water, nutrients, and sunlight, affecting the quality and productivity of the desired crops. Weed detection is crucial for the application of precision agriculture methods and for this purpose machine learning techniques can be used, specifically convolutional neural networks (CNN). This study focuses on the search for CNN architectures and technology used to detect and identify weeds in different crops; 61 articles applying CNN architectures and technology were analyzed in the last five years (2019-2023). The results show the used of different devices to acquire the image for training, such as digital cameras, smartphones, and drone cameras. Additionally, the YOLO family and algorithms are the most widely adopted architectures, followed by VGG, ResNet, Faster R-CNN, AlexNet, and MobileNet, respectively. This study provides an update on CNNs that will serve as a starting point for researchers wishing to implement these weed detection and identification techniques.

This article presents the results obtained in the evaluation of the LoRa technology as an alternative communication protocol for agricultural applications, by making measurements in situ in a rural area between the municipalities of Sibaté and Granada in Cundinamarca, Colombia. The device for testing was selected using a document search, thus choosing the TTGO esp32 board. Additionally, a temperature sensor and a variable resistance were included to obtain measurements from a transmitter node whose data was sent using the LoRa protocol to a receiver node with a distance of 500 meters between them and an approximate elevation gradient of 4° difference between the points. Finally, the information collected in the tests was subjected to a descriptive statistical analysis that allowed establishing the most reliable parameters of the protocol and identifying the appropriate spreading factor and bandwidth.

Water stress due to soil water deficit is one of the limitations in the soybean production, which can be detected with multivariate statistical analysis and spectral reflectance signals, in the visible and near infrared range. This work was conducted to determine a spectral pattern during the stages of plant development from three conditions of soil water content. Cross validation was used for validation of the classification model, with an accuracy of 82.5%, and the model reached a mean of 82 and 90% of sensitivity and specificity, respectively, at the phenological state of pod filling. The regions with the highest correlation between factors and wavelengths were located at 400-600 nm and 1850-2100 nm, which are related with the peaks of water energy absorbance associated to the hydric state of the plant