ABSTRACT The advent of unmanned aerial vehicles (UAVs) has significantly advanced precision agriculture by providing scalable, high-resolution imagery for farmland monitoring. However, the generation of descriptive captions from UAV images for agricultural applications remains an underexplored area of research. This study addresses this gap by introducing a prioritized retrieval model designed to generate captions that characterize crop growth stages from aerial imagery acquired by UAVs. In the absence of an existing benchmark, a dedicated Crop Growth Image Captions Dataset was developed to facilitate experimentation. The quality of the captions produced by the proposed model, in comparison with contemporary methods, was assessed using both objective evaluation metrics and subjective criteria. By leveraging UAV-based imagery, the proposed approach contributes towards automation in crop assessment, thereby mitigating the limitations of conventional, labour-intensive techniques and enhancing efficiency and accuracy in agricultural monitoring.

The use of unmanned aerial vehicles (UAVs) has emerged as a promising tool to maximize agricultural productivity and is useful in precision and sustainable insect pest control, with less impact on the environment and human health. In this review, we revealed how UAVs are valuable in dealing with pesticide resistance issues, reducing the ecological footprint of practices during pest management, and how they are safer for natural enemies. These UAVs have the ability to considerably reduce the overall dosage of insecticides with targeted delivery, which minimizes the risk of resistance development in insect pests against tested pesticides. Moreover, the addition of advanced technologies, such as computer vision systems, allows UAVs to optimize pesticide usage based on real-time data. The delivery of biological control agents by UAVs further supports ecological sustainability. However, a careful consideration is needed on the impact of UAV-applied chemicals on non-target organisms and ecosystems. To minimize drift, strategies such as optimized flight parameters, use of suitable nozzles, and integration of adjuvants should be adopted, which are important to ensure environmental safety. Future research on UAV technology, application methods, pesticide formulation, and natural enemy delivery is crucial to maximize the benefits of this innovative technique and ensure the safety of agroecosystems. © 2026 Society of Chemical Industry.

DUST: A framework for quantifying dugong-seagrass interactions using low-cost UAVs.

With characteristics such as low cost, flexibility, and scalability, unmanned aerial vehicle (UAV) swarms demonstrate superior performance over single unmanned platforms and manned aircraft in search and surveillance missions. However, the conflicts between individual decisions and the tradeoff between search and connectivity in limited sensing areas still render UAV swarm search and surveillance inherently challenging. This paper proposes a subgroup differentiation-based decision-making framework for the UAV swarm, where two kinds of roles (informers and relays) are considered and each UAV can autonomously switch its role according to the task demand and environmental changes. The relay nodes provide larger communication scopes for connectivity preservation and contribute to the relaxation of informers’ constrained actions. In this way, the informing UAVs can maintain well-preserved explorations during the search process. The subgroup differentiation is based on a distributed framework where two sequentially linked auction-based operators are respectively developed for the action policies of informers and relays. The impact time control guidance is used for simultaneous arrival to realize the synchronous information fusion of swarm search findings. Simulation results demonstrate the efficient explorations, less conservativeness, and higher coverage efficiencies of the proposed method over existing advanced approaches in confined sensing environments.

Existing forest smoke detection models face limitations in recognizing small targets, achieving real-time performance, and maintaining high inference efficiency on edge devices. To overcome these challenges, this study proposes a novel lightweight multi-scale You Only Look Once (LM-YOLO) model that enhances detection accuracy and real-time capability while ensuring computational efficiency. The LM-YOLO integrates the backbone feature extraction and multi-scale fusion mechanisms of YOLOv12n. Specifically, we introduce the RFMBlock to enhance the fused representation of shallow and deep features, considering the visual characteristics of forest smoke. Meanwhile, the two-path downsampling module achieves efficient downsampling with minimal spatial information loss. To further improve localization accuracy for small and blurred smoke regions, we employ the Shape-IoU loss function. Extensive experiments on the public try123-v4 forest fire smoke dataset demonstrate that, compared with the baseline YOLOv12n under the same experimental conditions, LM-YOLO reduces the number of parameters by 34.5%, decreases the computational cost by 30.2%, and improves detection precision by 4.7%. In addition, the model achieves 92.7% precision on the public try7 dataset. These results outperform existing methods and highlight the strong adaptability of LM-YOLO for edge deployment, ultimately providing reliable technical support for early forest fire warning using unmanned aerialvehicles.

Regarding the dependence of the traditional extended Kalman filter (EKF) algorithm on Gaussian measurement noise for identifying aerodynamic parameters in unmanned aerial vehicles (UAVs), an interacting noise model–based extended Kalman filter (INM-EKF) is proposed. Firstly, the longitudinal aerodynamic parameters of the unmanned aerial vehicle are selected as the identification object, and the identification system is modeled accordingly. The non-Gaussian measurement noise is approximated using a Gaussian mixture model. Subsequently, the expectation-maximization algorithm is employed to estimate the parameters of this model. Filtering calculations are then performed for multiple identification systems that incorporate Gaussian noise. The estimated values of the aerodynamic parameters are obtained by fusing the calculated results from each model. Finally, simulation experiments were conducted to compare the performance of the proposed method with three existing filtering techniques in environments characterized by Gaussian mixture noise and α-stable noise. The results indicate that the INM-EKF method outperforms the other techniques in terms of identification accuracy and convergence speed, significantly enhancing the accuracy of aerodynamic parameter identification for unmanned aerial vehicles.

Operational evaluation of unmanned aerial vehicle-applied granular larvicides in an integrated mosquito management program.

• A novel BIM-AKF framework with AAS support is proposed for the INS VC stabilization. • VC errors are modeled using scalar GM/sinusoidal stochastic processes via ACF/AV tools. • The proposed approach is experimentally validated using real UAV flight navigation data. Vertical Channel (VC) instability poses a critical challenge in Inertial Navigation Systems (INSs), particularly for aerial applications. This work proposes a novel hybrid architecture that integrates Baro-Inertial Mechanization (BIM), as a pre-stabilizer for the INS VC, with an Augmented Kalman Filter (AKF) that models only the residual VC errors after BIM correction. Such an architecture is therefore referred to as BIM-AKF. Unlike standard Kalman Filters (KFs), which embed the full unstable INS VC dynamics, BIM-AKF leverages either Auto-Correlation Function (ACF) or Allan Variance (AV) analysis of BIM-only outputs to optimally parameterize residual error propagation in state-space. The framework seamlessly supports Additional Aiding Sensors (AASs), e.g., Global Navigation Satellite Systems (GNSSs). Real Unmanned Aerial Vehicle (UAV) flights with intentional GNSS outages demonstrate that the proposed BIM-AKF outperforms existing methods, establishing a new benchmark in robust multi-sensor integrated navigation.

Ground vibration testing (GVT) plays a key role in the validation of numerical models and the assessment of aeroelastic stability in lightweight aircraft structures. This study presents an experimental and numerical investigation of a full-scale composite flying wing unmanned aerial vehicle (UAV) intended for vertical take-off and landing operations. Due to its low structural mass and highly integrated configuration, the aircraft exhibits increased sensitivity to modeling assumptions, boundary conditions, and measurement uncertainties. A finite element model was developed in Ansys, incorporating detailed laminate definitions and the internal sandwich structure. Experimental modal testing was performed under free-free boundary conditions using an electrodynamic shaker and a distributed measurement consisting of 94 response locations. Frequency Response Functions (FRFs), coherence analysis, and the Complex Mode Indication Function (CMIF) were employed to identify the dominant structural modes. Particular attention was given to the bending and torsional modes that govern aeroelastic behavior. Comparison of experimental and numerical results showed good agreement in mode shapes, while discrepancies in natural frequencies ranged from 10.4% to 20.1%. The results demonstrate that the model adequately captures the dynamic behavior of the aircraft and provides a reliable basis for future aeroelastic and flutter analyses of lightweight composite flying wing.

Observer-based prescribed-time optimal neural consensus control for six-rotor UAVs: A novel actor-critic reinforcement learning strategy.

MACTrack: Spatiotemporal context propagation with motion compensation for anti-UAV tracking.

DRONet: occlusion-mastering multi-object detection tailored for unmanned aerial vehicles

AUV-Net: An adaptive feature fusion network for small object detection in unmanned aerial vehicle scenes

Context-aware and deformation-adaptive small unmanned aerial vehicles detection via parallel attention and multi-scale fusion

Accurate geolocalization of ground targets from unmanned aerial vehicles (UAVs) is critically limited by pose estimation errors and the scarcity of active ranging sensors. To address these challenges, we propose a pipeline that integrates reference image cropping, robust cross-view matching, and geographic projection to estimate real-world coordinates using 2.5D reference maps. For evaluation, we introduce SkyPin, the first large-scale benchmark of its kind, designed to comprehensively test UAV-based localization methods. It comprises UAV imagery from eight diverse environments, featuring both visible and thermal infrared modalities under a wide range of conditions, including variations in weather, time of day, flight altitude, and camera perspective. All ground targets are annotated with centimeter-accuracy Real-Time Kinematic (RTK) coordinates. We establish a comprehensive benchmark by evaluating a series of feature matching methods combined with different projection strategies, allowing systematic comparison of algorithm performance. Representative results show that RoMa combined with PnP-based raytracing achieves the best overall performance, reaching a median 2D error of 0.87 m and Recall@5m values of 0.94 and 0.98 on RGB and thermal infrared UAV-map settings, respectively. Further analysis reveals that performance degrades in challenging mountainous scenes and under large viewing-angle variations, highlighting terrain relief and UAV perspective changes as remaining critical challenges for robust target geo-localization. The full dataset and implementation code will be made publicly available to facilitate future research in UAV-based geolocalization.

The combination of unmanned aerial vehicle technology and deep learning has revolutionized precision agriculture by facilitating the monitoring of crops using UAVs. The current pest and weed detection methods have limitations in that they use individual neural networks for pest and weed detection, leading to methodological fragmentation. This new framework clearly explains about a new pest and weed detection system dubbed AgriYOLO12-Dual, based on a unified detection framework. The framework employs YOLOv12, the initial YOLO variant to use self-attention as a basic computing unit. The attention-centric YOLOv12 framework system has a dual-branch encoder network that takes RGB and multispectral images from a UAV and passes them separately to a neural network. The outputs of the networks are then combined at a cross-modal fusion point using Area Attention. The attention-centric YOLOv12 framework system has three key innovations. The primary one is the Area Attention mechanism, it has a large receptive field and linear computational cost. The second innovation is the use of Residual Efficient Layer Aggregation Networks (R-ELAN), that allows for the training of large attention models. The third innovation is the use of Flash Attention to reduce memory usage by 38%. The multimodal dual-branch attention-centric YOLOv12 framework was trained on 15,000 annotated images of maize, soybean, and wheat crops at various stages of growth. The results showed that the attention-centric YOLOv12 framework system achieved a weed detection mAP of 90.1% and a pest detection mAP of 93.4%, with an inference time of 28.7ms on a Jetson Orin device. The small target recall was improved from 67.4% to 79.3%, and the low-light detection mAP was improved from 76.2% to 87.2%. The results of the experiments showed that attention-based models have a significant improvement in pest and weed detection accuracy without additional complexity.

Woody plant encroachment (WPE) is widespread in savanna and wetland ecosystems, yet its effects on phenological dynamics at fine spatial scales remain poorly quantified. Here, we assessed how woody plant encroachment alters vegetative and reproductive phenology in a Vereda palm swamp ecosystem of the Brazilian Cerrado by integrating high-resolution unmanned aerial vehicle (RPAS; Remotely Piloted Aircraft System) imagery with ground-based phenological monitoring. Ten regions of interest representing WPE and non-invaded native Vereda vegetation were monitored monthly for twelve consecutive months using drone imagery, with 1 × 1m sampling grids (≈ 260m² per region). Vegetation greenness was quantified from monthly orthomosaics using the Green Chromatic Coordinate (GCC) and analyzed with linear mixed-effects models including vegetation type, time, and their interaction. Flowering phenology was monitored biweekly from January to December 2023 in transects located in the middle region of the same Vereda area. Three WPE plant species and the native Vereda community occurring in the same area were analyzed using circular statistics and permutation tests. Woody plant encroachment areas exhibited significantly higher GCC values than native Vereda vegetation, indicating greater and more persistent greenness throughout the year. A significant vegetation type × time interaction revealed distinct phenological trajectories, with WPE areas maintaining elevated greenness during dry-wet seasonal transitions and possibly resulting in greater evapotranspiration and water use during the end of the dry season. Flowering phenology of WPE species was highly synchronized (mean angle = 3.45rad; r = 0.65), whereas native Vereda species showed a more diffuse flowering pattern (mean angle = 1.99rad; r = 0.27). Although a temporal flowering peak shift of approximately three months was observed between groups, this difference was not statistically significant. Overall, woody plant encroachment substantially modifies vegetative phenological dynamics and tends to homogenize flowering timing in Vereda ecosystems, highlighting the value of integrating RPAS-based monitoring and field observations to detect functional changes in Cerrado wetlands.

The article presents the results of the analysis of the use of geographic information systems (GIS) in the management of agricultural land resources in the Omsk region. A list of organizations using GIS is identified, including Rosreestr, Roskadastr, the Ministry of Agriculture, the Agrochemcenter and private geodetic enterprises. The dynamics of the number of organizations using GIS for 2020–2024 is shown, which decreased by 3.1% due to economic factors, but the share of farms using open-source software is growing. The land management structure, including federal, regional and municipal bodies, is described, and key problems are identified: departmental disunity, outdated cartographic materials, lack of field measurements and low data integration. Solutions are proposed, including the creation of a unified geoinformation environment based on the integration of remote sensing data, unmanned aerial vehicles, IoT sensors and field surveys. The expected effectiveness of the proposed technology will increase the efficiency and reliability of information, reduce land management costs and improve control over the use of agricultural land.

How are Unmanned Aerial Vehicles (UAVs) Revolutionizing Forest Operations? A Systematic Review of Current Applications and Future Opportunities.

This study addresses the need for efficient aerodynamic design in fixed-wing unmanned aerial vehicles (UAVs) for aerial mapping applications, where flight stability and cruising performance are critical. The research aims to optimize wing geometry parameters to achieve the desired cruising speed while minimizing drag and ensuring static stability. The methodology integrates conceptual and preliminary design approaches, followed by aerodynamic simulations using XFLR-5 with the vortex lattice methodology (VLM-2). Three design variables, winglet length, cant angle, and twist angle, are systematically varied, and the response surface method (RSM) is employed to model and optimize their effects on lift, drag, and airspeed. The optimization results indicate that the optimal configuration achieves a cruising speed of 16.8 m/s with improved lift characteristics (CL ≈ 0.48) and controlled drag (CD ≈ 0.022). Further analysis confirms that the optimized UAV satisfies longitudinal, lateral, and directional static stability criteria under various control surface deflections. In conclusion, the integration of RSM with aerodynamic simulation provides an effective and systematic framework for enhancing the UAV performance and stability, particularly in aerial mapping missions.