RESUMEN

This study leverages the BERTopic algorithm to analyze the evolution of research within precision agriculture, identifying 37 distinct topics categorized into eight subfields: Data Analysis, IoT, UAVs, Soil and Water Management, Crop and Pest Management, Livestock, Sustainable Agriculture, and Technology Innovation. By employing BERTopic, based on a transformer architecture, this research enhances topic refinement and diversity, distinguishing it from traditional reviews. The findings highlight a significant shift towards IoT innovations, such as security and privacy, reflecting the integration of smart technologies with traditional agricultural practices. Notably, this study introduces a comprehensive popularity index that integrates trend intensity with topic proportion, providing nuanced insights into topic dynamics across countries and journals. The analysis shows that regions with robust research and development, such as the USA and Germany, are advancing in technologies like Machine Learning and IoT, while the diversity in research topics, assessed through information entropy, indicates a varied global research scope. These insights assist scholars and research institutions in selecting research directions and provide newcomers with an understanding of the field's dynamics.

RESUMEN

In this study, a deep learning method combining knowledge graph and diffusion Transformer has been proposed for cucumber disease detection. By incorporating the diffusion attention mechanism and diffusion loss function, the research aims to enhance the model's ability to recognize complex agricultural disease features and to address the issue of sample imbalance efficiently. Experimental results demonstrate that the proposed method outperforms existing deep learning models in cucumber disease detection tasks. Specifically, the method achieved a precision of 93%, a recall of 89%, an accuracy of 92%, and a mean average precision (mAP) of 91%, with a frame rate of 57 frames per second (FPS). Additionally, the study successfully implemented model lightweighting, enabling effective operation on mobile devices, which supports rapid on-site diagnosis of cucumber diseases. The research not only optimizes the performance of cucumber disease detection, but also opens new possibilities for the application of deep learning in the field of agricultural disease detection.

RESUMEN

WeedCube dataset consists of hyperspectral images of three crops (canola, soybean, and sugarbeet) and four invasive weeds species (kochia, common waterhemp, redroot pigweed, and common ragweed). Plants were grown in two separate greenhouses and plant canopies were captured from a top-down camera angle. A push-broom hyperspectral sensor in the visible near infrared region of 400-1000 nm was used for data collection. The dataset includes 160 calibrated images. The number of images can be further increased by selection of smaller region of interests (ROIs). Dataset is supplemented by Jupyter Notebook scripts that help in data augmentation, spectral pre-processing, ROI selection for points and images, and data visualization. The primary purpose of this dataset is to support weed classification or identification studies by enhancing existing training datasets and validating the generalization capabilities of existing models. Owing to the three-dimensional (3D) nature of hyperspectral images, this dataset can also be utilized by researchers and educators across various domains for the development and testing of deep learning algorithms, the creation of automated data processing pipelines effective for 3D data, the development of tools for 3D data visualization, the creation of innovative solutions for data compression, and addressing system memory issues associated with high-dimensional data.

RESUMEN

Precision weed management (PWM), driven by machine vision and deep learning (DL) advancements, not only enhances agricultural product quality and optimizes crop yield but also provides a sustainable alternative to herbicide use. However, existing DL-based algorithms on weed detection are mainly developed based on supervised learning approaches, typically demanding large-scale datasets with manual-labeled annotations, which can be time-consuming and labor-intensive. As such, label-efficient learning methods, especially semi-supervised learning, have gained increased attention in the broader domain of computer vision and have demonstrated promising performance. These methods aim to utilize a small number of labeled data samples along with a great number of unlabeled samples to develop high-performing models comparable to the supervised learning counterpart trained on a large amount of labeled data samples. In this study, we assess the effectiveness of a semi-supervised learning framework for multi-class weed detection, employing two well-known object detection frameworks, namely FCOS (Fully Convolutional One-Stage Object Detection) and Faster-RCNN (Faster Region-based Convolutional Networks). Specifically, we evaluate a generalized student-teacher framework with an improved pseudo-label generation module to produce reliable pseudo-labels for the unlabeled data. To enhance generalization, an ensemble student network is employed to facilitate the training process. Experimental results show that the proposed approach is able to achieve approximately 76% and 96% detection accuracy as the supervised methods with only 10% of labeled data in CottonWeedDet3 and CottonWeedDet12, respectively. We offer access to the source code (https://github.com/JiajiaLi04/SemiWeeds), contributing a valuable resource for ongoing semi-supervised learning research in weed detection and beyond.

RESUMEN

This study focuses on semantic segmentation in crop Opuntia spp. orthomosaics; this is a significant challenge due to the inherent variability in the captured images. Manual measurement of Opuntia spp. vegetation areas can be slow and inefficient, highlighting the need for more advanced and accurate methods. For this reason, we propose to use deep learning techniques to provide a more precise and efficient measurement of the vegetation area. Our research focuses on the unique difficulties posed by segmenting high-resolution images exceeding 2000 pixels, a common problem in generating orthomosaics for agricultural monitoring. The research was carried out on a Opuntia spp. cultivation located in the agricultural region of Tulancingo, Hidalgo, Mexico. The images used in this study were obtained by drones and processed using advanced semantic segmentation architectures, including DeepLabV3+, UNet, and UNet Style Xception. The results offer a comparative analysis of the performance of these architectures in the semantic segmentation of Opuntia spp., thus contributing to the development and improvement of crop analysis techniques based on deep learning. This work sets a precedent for future research applying deep learning techniques in agriculture.

RESUMEN

Spectral signatures allow the characterization of a surface from the reflected or emitted energy along the electromagnetic spectrum. This type of measurement has several potential applications in precision agriculture. However, capturing the spectral signatures of plants requires specialized instruments, either in the field or the laboratory. The cost of these instruments is high, so their incorporation in crop monitoring tasks is not massive, given the low investment in agricultural technology. This paper presents a low-cost clamp to capture spectral leaf signatures in the laboratory and the field. The clamp can be 3D printed using PLA (polylactic acid); it allows the connection of 2 optical fibers: one for a spectrometer and one for a light source. It is designed for ease of use and holds a leave firmly without causing damage, allowing data to be collected with less disturbance. The article compares signatures captured directly using a fiber and the proposed clamp; noise reduction across the spectrum is achieved with the clamp.

RESUMEN

The fraction of absorbed photosynthetically active radiation (FAPAR) and the photosynthesis rate (Pn) of maize canopies were identified as essential photosynthetic parameters for accurately estimating vegetation growth and productivity using multispectral vegetation indices (VIs). Despite their importance, few studies have compared the effectiveness of multispectral imagery and various machine learning techniques in estimating these photosynthetic traits under high vegetation coverage. In this study, seventeen multispectral VIs and four machine learning (ML) algorithms were utilized to determine the most suitable model for estimating maize FAPAR and Pn during the kharif and rabi seasons at Tamil Nadu Agricultural University, Coimbatore, India. Results demonstrate that indices such as OSAVI, SAVI, EVI-2, and MSAVI-2 during the kharif and MNDVIRE and MSRRE during the rabi season outperformed others in estimating FAPAR and Pn values. Among the four ML methods of random forest (RF), extreme gradient boosting (XGBoost), support vector regression (SVR), and multiple linear regression (MLR) considered, RF consistently showed the most effective fitting effect and XGBoost demonstrated the least fitting accuracy for FAPAR and Pn estimation. However, SVR with R2 = 0.873 and RMSE = 0.045 during the kharif and MLR with R2 = 0.838 and RMSE = 0.053 during the rabi season demonstrated higher fitting accuracy, particularly notable for FAPAR prediction. Similarly, in the prediction of Pn, MLR showed higher fitting accuracy with R2 = 0.741 and RMSE = 2.531 during the kharif and R2 = 0.955 and RMSE = 1.070 during the rabi season. This study demonstrated the potential of combining UAV-derived VIs with ML to develop accurate FAPAR and Pn prediction models, overcoming VI saturation in dense vegetation. It underscores the importance of optimizing these models to improve the accuracy of maize vegetation assessments during various growing seasons.

RESUMEN

In strawberry cultivation, precise disease management is crucial for maximizing yields and reducing unnecessary fungicide use. Traditional methods for measuring leaf wetness duration (LWD), a critical factor in assessing the risk of fungal diseases such as botrytis fruit rot and anthracnose, have been reliant on sensors with known limitations in accuracy and reliability and difficulties with calibrating. To overcome these limitations, this study introduced an innovative algorithm for leaf wetness detection systems employing high-resolution imaging and deep learning technologies, including convolutional neural networks (CNNs). Implemented at the University of Florida's Plant Science Research and Education Unit (PSREU) in Citra, FL, USA, and expanded to three additional locations across Florida, USA, the system captured and analyzed images of a reference plate to accurately determine the wetness and, consequently, the LWD. The comparison of system outputs with manual observations across diverse environmental conditions demonstrated the enhanced accuracy and reliability of the artificial intelligence-driven approach. By integrating this system into the Strawberry Advisory System (SAS), this study provided an efficient solution to improve disease risk assessment and fungicide application strategies, promising significant economic benefits and sustainability advances in strawberry production.

Asunto(s)

RESUMEN

Biological nitrogen fixation (BNF) by symbiotic bacteria plays a vital role in sustainable agriculture. However, current quantification methods are often expensive and impractical. This study explores the potential of Raman spectroscopy, a non-invasive technique, for rapid assessment of BNF activity in soybeans. Raman spectra were obtained from soybean plants grown with and without rhizobia bacteria to identify spectral signatures associated with BNF. δN15 isotope ratio mass spectrometry (IRMS) was used to determine actual BNF percentages. Partial least squares regression (PLSR) was employed to develop a model for BNF quantification based on Raman spectra. The model explained 80% of the variation in BNF activity. To enhance the model's specificity for BNF detection regardless of nitrogen availability, a subsequent elastic net (Enet) regularisation strategy was implemented. This approach provided insights into key wavenumbers and biochemicals associated with BNF in soybeans.

Asunto(s)

RESUMEN

In both plant breeding and crop management, interpretability plays a crucial role in instilling trust in AI-driven approaches and enabling the provision of actionable insights. The primary objective of this research is to explore and evaluate the potential contributions of deep learning network architectures that employ stacked LSTM for end-of-season maize grain yield prediction. A secondary aim is to expand the capabilities of these networks by adapting them to better accommodate and leverage the multi-modality properties of remote sensing data. In this study, a multi-modal deep learning architecture that assimilates inputs from heterogeneous data streams, including high-resolution hyperspectral imagery, LiDAR point clouds, and environmental data, is proposed to forecast maize crop yields. The architecture includes attention mechanisms that assign varying levels of importance to different modalities and temporal features that, reflect the dynamics of plant growth and environmental interactions. The interpretability of the attention weights is investigated in multi-modal networks that seek to both improve predictions and attribute crop yield outcomes to genetic and environmental variables. This approach also contributes to increased interpretability of the model's predictions. The temporal attention weight distributions highlighted relevant factors and critical growth stages that contribute to the predictions. The results of this study affirm that the attention weights are consistent with recognized biological growth stages, thereby substantiating the network's capability to learn biologically interpretable features. Accuracies of the model's predictions of yield ranged from 0.82-0.93 R2 ref in this genetics-focused study, further highlighting the potential of attention-based models. Further, this research facilitates understanding of how multi-modality remote sensing aligns with the physiological stages of maize. The proposed architecture shows promise in improving predictions and offering interpretable insights into the factors affecting maize crop yields, while demonstrating the impact of data collection by different modalities through the growing season. By identifying relevant factors and critical growth stages, the model's attention weights provide valuable information that can be used in both plant breeding and crop management. The consistency of attention weights with biological growth stages reinforces the potential of deep learning networks in agricultural applications, particularly in leveraging remote sensing data for yield prediction. To the best of our knowledge, this is the first study that investigates the use of hyperspectral and LiDAR UAV time series data for explaining/interpreting plant growth stages within deep learning networks and forecasting plot-level maize grain yield using late fusion modalities with attention mechanisms.

RESUMEN

The pollination process of kiwifruit flowers plays a crucial role in kiwifruit yield. Achieving accurate and rapid identification of the four stages of kiwifruit flowers is essential for enhancing pollination efficiency. In this study, to improve the efficiency of kiwifruit pollination, we propose a novel full-stage kiwifruit flower pollination detection algorithm named KIWI-YOLO, based on the fusion of frequency-domain features. Our algorithm leverages frequency-domain and spatial-domain information to improve recognition of contour-detailed features and integrates decision-making with contextual information. Additionally, we incorporate the Bi-Level Routing Attention (BRA) mechanism with C3 to enhance the algorithm's focus on critical areas, resulting in accurate, lightweight, and fast detection. The algorithm achieves a m A P 0.5 of 91.6% with only 1.8M parameters, the AP of the Female class and the Male class reaches 95% and 93.5%, which is an improvement of 3.8%, 1.2%, and 6.2% compared with the original algorithm. Furthermore, the Recall and F1-score of the algorithm are enhanced by 5.5% and 3.1%, respectively. Moreover, our model demonstrates significant advantages in detection speed, taking only 0.016s to process an image. The experimental results show that the algorithmic model proposed in this study can better assist the pollination of kiwifruit in the process of precision agriculture production and help the development of the kiwifruit industry.

RESUMEN

Sensors used in precision agriculture for the detection of heavy metals in irrigation water are generally expensive and sometimes their deployment and maintenance represent a permanent investment to keep them in operation, leaving a lasting polluting footprint in the environment at the end of their lifespan. This represents an area of opportunity to design new biological devices that can replace part, or all of the sensors currently used. In this article, a novel workflow is proposed to fully carry out the complete process of design, modeling, and simulation of reprogrammable microorganisms in silico. As a proof-of-concept, the workflow has been used to design three whole-cell biosensors for the detection of heavy metals in irrigation water, namely arsenic, mercury and lead. These biosensors are in compliance with the concentration limits established by the World Health Organization (WHO). The proposed workflow allows the design of a wide variety of completely in silico biodevices, which aids in solving problems that cannot be easily addressed with classical computing. The workflow is based on two technologies typical of synthetic biology: the design of synthetic genetic circuits, and in silico synthetic engineering, which allows us to address the design of reprogrammable microorganisms using software and hardware to develop theoretical models. These models enable the behavior prediction of complex biological systems. The output of the workflow is then exported in the form of complete genomes in SBOL, GenBank and FASTA formats, enabling their subsequent in vivo implementation in a laboratory. The present proposal enables professionals in the area of computer science to collaborate in biotechnological processes from a theoretical perspective previously or complementary to a design process carried out directly in the laboratory by molecular biologists. Therefore, key results pertaining to this work include the fully in silico workflow that leads to designs that can be tested in the lab in vitro or in vivo, and a proof-of-concept of how the workflow generates synthetic circuits in the form of three whole-cell heavy metal biosensors that were designed, modeled and simulated using the workflow. The simulations carried out show realistic spatial distributions of biosensors reacting to different concentrations (zero, low and threshold level) of heavy metal presence and at different growth phases (stationary and exponential) that are backed up by the whole design and modeling phases of the workflow.

RESUMEN

Winter is the season of main concern for beekeepers since the temperature, humidity, and potential infection from mites and other diseases may lead the colony to death. As a consequence, beekeepers perform invasive checks on the colonies, exposing them to further harm. This paper proposes a novel design of an instrumented beehive involving color cameras placed inside the beehive and at the bottom of it, paving the way for new frontiers in beehive monitoring. The overall acquisition system is described focusing on design choices towards an effective solution for internal, contactless, and stress-free beehive monitoring. To validate our approach, we conducted an experimental campaign in 2023 and analyzed the collected images with YOLOv8 to understand if the proposed solution can be useful for beekeepers and what kind of information can be derived from this kind of monitoring, including the presence of Varroa destructor mites inside the beehive. We experimentally found that the observation point inside the beehive is the most challenging due to the frequent movements of the bees and the difficulties related to obtaining in-focus images. However, from these images, it is possible to find Varroa destructor mites. On the other hand, the observation point at the bottom of the beehive showed great potential for understanding the overall activity of the colony.

Asunto(s)

RESUMEN

(1) Background: Yield-monitoring systems are widely used in grain crops but are less advanced for hay and forage. Current commercial systems are generally limited to weighing individual bales, limiting the spatial resolution of maps of hay yield. This study evaluated an Uncrewed Aerial Vehicle (UAV)-based imaging system to estimate hay yield. (2) Methods: Data were collected from three 0.4 ha plots and a 35 ha hay field of red clover and timothy grass in September 2020. A multispectral camera on the UAV captured images at 30 m (20 mm pixel-1) and 50 m (35 mm pixel-1) heights. Eleven Vegetation Indices (VIs) and five texture features were calculated from the images to estimate biomass yield. Multivariate regression models (VIs and texture features vs. biomass) were evaluated. (3) Results: Model R2 values ranged from 0.31 to 0.68. (4) Conclusions: Despite strong correlations between standard VIs and biomass, challenges such as variable image resolution and clarity affected accuracy. Further research is needed before UAV-based yield estimation can provide accurate, high-resolution hay yield maps.

Asunto(s)

RESUMEN

Precision agriculture has revolutionized crop management and agricultural production, with LiDAR technology attracting significant interest among various technological advancements. This extensive review examines the various applications of LiDAR in precision agriculture, with a particular emphasis on its function in crop cultivation and harvests. The introduction provides an overview of precision agriculture, highlighting the need for effective agricultural management and the growing significance of LiDAR technology. The prospective advantages of LiDAR for increasing productivity, optimizing resource utilization, managing crop diseases and pesticides, and reducing environmental impact are discussed. The introduction comprehensively covers LiDAR technology in precision agriculture, detailing airborne, terrestrial, and mobile systems along with their specialized applications in the field. After that, the paper reviews the several uses of LiDAR in agricultural cultivation, including crop growth and yield estimate, disease detection, weed control, and plant health evaluation. The use of LiDAR for soil analysis and management, including soil mapping and categorization and the measurement of moisture content and nutrient levels, is reviewed. Additionally, the article examines how LiDAR is used for harvesting crops, including its use in autonomous harvesting systems, post-harvest quality evaluation, and the prediction of crop maturity and yield. Future perspectives, emergent trends, and innovative developments in LiDAR technology for precision agriculture are discussed, along with the critical challenges and research gaps that must be filled. The review concludes by emphasizing potential solutions and future directions for maximizing LiDAR's potential in precision agriculture. This in-depth review of the uses of LiDAR gives helpful insights for academics, practitioners, and stakeholders interested in using this technology for effective and environmentally friendly crop management, which will eventually contribute to the development of precision agricultural methods.

Asunto(s)

RESUMEN

Indoor production of basil (Ocimum basilicum L.) is influenced by light spectrum, photosynthetic photon flux density (PPFD), and the photoperiod. To investigate the effects of different lighting on growth, chlorophyll content, and secondary metabolism, basil plants were grown from seedlings to fully expanded plants in microcosm devices under different light conditions: (a) white light at 250 and 380 µmol·m-2·s-1 under 16/8 h light/dark and (b) white light at 380 µmol·m-2·s-1 under 16/8 and 24/0 h light/dark. A higher yield was recorded under 380 µmol·m-2·s-1 compared to 250 µmol·m-2·s-1 (fresh and dry biomasses 260.6 ± 11.3 g vs. 144.9 ± 14.6 g and 34.1 ± 2.6 g vs. 13.2 ± 1.4 g, respectively), but not under longer photoperiods. No differences in plant height and chlorophyll content index were recorded, regardless of the PPFD level and photoperiod length. Almost the same volatile organic compounds (VOCs) were detected under the different lighting treatments, belonging to terpenes, aldehydes, alcohols, esters, and ketones. Linalool, eucalyptol, and eugenol were the main VOCs regardless of the lighting conditions. The multivariate data analysis showed a sharp separation of non-volatile metabolites in apical and middle leaves, but this was not related to different PPFD levels. Higher levels of sesquiterpenes and monoterpenes were detected in plants grown under 250 µmol·m-2·s-1 and 380 µmol·m-2·s-1, respectively. A low separation of non-volatile metabolites based on the photoperiod length and VOC overexpression under longer photoperiods were also highlighted.

RESUMEN

Detecting volatile organic compounds (VOCs) emitted from different plant species and their organs can provide valuable information about plant health and environmental factors that affect them. For example, limonene emission can be a biomarker to monitor plant health and detect stress. Traditional methods for VOC detection encounter challenges, prompting the proposal of novel approaches. In this study, we proposed integrating electrospinning, molecular imprinting, and conductive nanofibers to fabricate limonene sensors. In detail, polyvinylpyrrolidone (PVP) and polyacrylic acid (PAA) served here as fiber and cavity formers, respectively, with multiwalled carbon nanotubes (MWCNT) enhancing conductivity. We developed one-step monolithic molecularly imprinted fibers, where S(-)-limonene was the target molecule, using an electrospinning technique. The functional cavities were fixed using the UV curing method, followed by a target molecule washing. This procedure enabled the creation of recognition sites for limonene within the nanofiber matrix, enhancing sensor performance and streamlining manufacturing. Humidity was crucial for sensor working, with optimal conditions at about 50% RH. The sensors rapidly responded to S(-)-limonene, reaching a plateau within 200 s. Enhancing fiber density improved sensor performance, resulting in a lower limit of detection (LOD) of 137 ppb. However, excessive fiber density decreased accessibility to active sites, thus reducing sensitivity. Remarkably, the thinnest mat on the fibrous sensors created provided the highest selectivity to limonene (Selectivity Index: 72%) compared with other VOCs, such as EtOH (used as a solvent in nanofiber development), aromatic compounds (toluene), and two other monoterpenes (α-pinene and linalool) with similar structures. These findings underscored the potential of the proposed integrated approach for selective VOC detection in applications such as precision agriculture and environmental monitoring.

RESUMEN

Soybean is an essential crop to fight global food insecurity and is of great economic importance around the world. Along with genetic improvements aimed at boosting yield, soybean seed composition also changed. Since conditions during crop growth and development influences nutrient accumulation in soybean seeds, remote sensing offers a unique opportunity to estimate seed traits from the standing crops. Capturing phenological developments that influence seed composition requires frequent satellite observations at higher spatial and spectral resolutions. This study introduces a novel spectral fusion technique called multiheaded kernel-based spectral fusion (MKSF) that combines the higher spatial resolution of PlanetScope (PS) and spectral bands from Sentinel 2 (S2) satellites. The study also focuses on using the additional spectral bands and different statistical machine learning models to estimate seed traits, e.g., protein, oil, sucrose, starch, ash, fiber, and yield. The MKSF was trained using PS and S2 image pairs from different growth stages and predicted the potential VNIR1 (705 nm), VNIR2 (740 nm), VNIR3 (783 nm), SWIR1 (1610 nm), and SWIR2 (2190 nm) bands from the PS images. Our results indicate that VNIR3 prediction performance was the highest followed by VNIR2, VNIR1, SWIR1, and SWIR2. Among the seed traits, sucrose yielded the highest predictive performance with RFR model. Finally, the feature importance analysis revealed the importance of MKSF-generated vegetation indices from fused images.

Asunto(s)

RESUMEN

Jataí is a pollinator of some crops; therefore, its sustainable management guarantees quality in the ecosystem services provided and implementation in precision agriculture. We acquired videos of natural and artificial hives in urban and rural environments with a camera positioned at the hive entrance. In this way, we obtained videos of the entrance of several colonies for multiple bee tracking and removed images from the videos for bee detectors. This data, their respective labels, and metadata make up the dataset. The dataset displays potential for utilization in computer vision tasks such as comparative studies of deep learning models. They can also integrate intelligent monitoring systems for natural and artificial hives.

RESUMEN

Digital image datasets for Precision Agriculture (PA) still need to be available. Many problems in this field of science have been studied to find solutions, such as detecting weeds, counting fruits and trees, and detecting diseases and pests, among others. One of the main fields of research in PA is detecting different crop types with aerial images. Crop detection is vital in PA to establish crop inventories, planting areas, and crop yields and to have information available for food markets and public entities that provide technical help to small farmers. This work proposes public access to a digital image dataset for detecting green onion and foliage flower crops located in the rural area of Medellín City - Colombia. This dataset consists of 245 images with their respective labels: green onion (Allium fistulosum), foliage flowers (Solidago Canadensis and Aster divaricatus), and non-crop areas prepared for planting. A total of 4315 instances were obtained, which were divided into subsets for training, validation, and testing. The classes in the images were labeled with the polygon method, which allows training machine learning algorithms for detection using bounding boxes or segmentation in the COCO format.