Search Results (4,052)

Unmanned aerial vehicle (UAV) operational training is often constrained by high equipment costs, safety risks, and limited training venues. In addition, novice operators may experience stress and anxiety in safety-risk environments, adversely affecting attention, decision-making, and task performance. To overcome these challenges, this study developed a virtual UAV training system using Unity and C#. The system enables bidirectional command communication and real-time state synchronization between virtual and physical UAVs through UDP-based communication. A quasi-experimental design was employed to evaluate the system’s technical feasibility and training performance. Sixty university students without prior UAV experience were assigned to either the virtual or physical training group. Learning performance was assessed in terms of knowledge acquisition, flight performance, learning motivation, cognitive load, and technology acceptance. Results showed that both groups demonstrated significant improvements in knowledge and flight performance, with statistical/practical equivalence within a prespecified margin after training, indicating comparable learning achievement. The virtual training group obtained higher “Relevance” scores on the ARCS motivation scale but experienced greater extraneous cognitive load, whereas the physical training group showed higher technology acceptance. Overall, the proposed system provides a safe, flexible, and effective alternative to conventional physical UAV training. Future work may optimize the user interface to reduce extraneous cognitive load and enhance technology acceptance, improving its applicability in skills-oriented education and training. Full article

Remote sensing research increasingly relies on heterogeneous satellite, UAV, hyperspectral, multispectral, SAR, and environmental monitoring data to support land, urban, hydrological, and environmental applications. However, these data are often affected by sensor differences, spatial and temporal heterogeneity, missing observations, irregular sampling, noise, and non-stationary environmental processes. This systematic review was conducted within the context of the Romanian Hub for Artificial Intelligence (HRIA) project, which supports the development of strategic artificial intelligence technologies. The review synthesizes current research on adaptive neural networks for remote sensing and Earth observation, with particular attention to Liquid Neural Networks and related continuous-time neural models. A systematic search was conducted across IEEE Xplore, Scopus, Web of Science, ScienceDirect, SpringerLink, Wiley Online Library, Google Scholar, and reference lists. After duplicate removal, screening, and full-text assessment, 61 studies published between 2018 and 2026 were included in the qualitative synthesis. The findings show that adaptive neural networks have gained increasing attention after 2022 and are mainly applied to image-centered remote sensing tasks, including classification, mapping, object detection, segmentation, enhancement, and change detection. Most studies adapt established deep learning architectures through multi-scale processing, adaptive feature fusion, attention mechanisms, graph relationships, or task-specific refinement. Continuous-time models are used less frequently but are relevant for irregular observations and dynamic environmental processes. Liquid Neural Networks remain emerging, and current evidence suggests only preliminary, task-specific relevance for irregular, noisy, multimodal, and dynamic remote sensing applications. Full article

Accurate instance-level perception of aerial view traffic accident scenes is the foundation of accident investigation. However, existing accident datasets mainly support event-level video analysis or coarse spatial localization, and rarely provide pixel-level vehicle masks together with accident-involved labels. To address this gap, we construct the Drone-oriented Accident Recognition and Segmentation (DARS) dataset, an instance segmentation dataset specifically developed for aerial view traffic accident scenes. DARS contains 7603 images and 49,613 vehicle instances with six classes jointly defined by vehicle type and accident-involved status. Statistical analysis reveals the class distribution, scale variation, and accident-type composition of the dataset. We further introduce a Frequency-Guided Adaptive Downsampling (FGAD) method into YOLO26n-seg to improve hierarchical feature extraction. FGAD performs content-adaptive aggregation of spatial candidate features, while wavelet-derived frequency information guides candidate weight estimation and provides an additional residual pathway. On DARS, the proposed method achieves 52.37% recall, 51.24% mAP@50, 41.18% mAP@75, and 36.77% mAP@50–95, outperforming other methods in terms of instance segmentation accuracy. On a UAV-based Vehicle Segmentation Dataset (UVSD), it consistently improves over YOLO26n-seg, reaching 75.61%, 57.74%, and 44.60%, respectively. These results support DARS as an instance-level benchmark and demonstrate the applicability of FGAD to aerial traffic perception, providing a basis for fine-grained accident scene analysis in intelligent and sustainable transportation systems. Full article

Improving pesticide application efficiency in dense pepper canopies requires optimization of Unmanned Aerial Vehicle (UAV) operational parameters based on both droplet deposition and pest control performance. This study evaluated the effects of flight height and nominal droplet size setting on canopy deposition characteristics and pest control efficacy using a DJI T60 plant protection UAV under field conditions. Three flight heights (2, 3, and 4 m) and five nominal droplet size settings (100, 150, 200, 250, and 300 µm) were investigated. Droplet coverage, deposition density, canopy penetration rate, deposition uniformity, and corrected control efficacy were used as evaluation indicators, and TOPSIS was applied for comprehensive parameter optimization. The results showed that both flight height and nominal droplet size setting exert a certain influence on droplet deposition characteristics and pest control efficacy. Under the specific conditions tested in this study, the 3 m flight height achieved favorable overall deposition performance, while the 250 µm nominal setting exhibited the highest effective coverage among the tested droplet size treatments. Pest control efficacy showed a consistent trend with droplet deposition performance, and the 3 m + 250 µm nominal setting combination achieved the highest corrected control efficacy at 7 DAT. According to the TOPSIS evaluation, the combination of 3 m flight height and 250 µm nominal droplet size setting achieved the best overall performance under the experimental conditions. Because actual airborne droplet size spectra were not independently measured, the recommendation refers to the tested operational setting rather than a measured droplet size population. Full article

Facade debonding may lead to falling components and pose safety risks in dense urban environments, but infrared thermal responses may also arise from non-defect facade components and environmental conditions. Conventional facade inspection methods are often labor-intensive, hazardous, and difficult to integrate into digital maintenance workflows. To support safer and more efficient facade inspection and maintenance information management, this study develops an engineering-oriented inspection and management framework that integrates unmanned aerial vehicle infrared thermography, intelligent defect recognition, visual result verification, and defect information management. A UAV-based infrared data acquisition scheme was established, and a self-constructed dataset containing 1035 thermal images was developed for the detection of suspected facade debonding regions and common thermal interference sources, including windows, air-conditioning units, and signage. A lightweight detection model was embedded as the recognition engine of the framework to balance detection reliability and deployment efficiency under practical inspection conditions. Experimental results show that the proposed method achieved an mAP@0.5 of 87.8%, with 1.64 million parameters and 4.2 GFLOPs, indicating its potential for rapid preliminary facade screening under the tested computing configuration. Beyond model evaluation, an application platform was developed to support infrared image and video input, automatic detection, result visualization, statistical analysis, and defect record storage. The proposed framework demonstrates the potential of combining UAV infrared inspection and digital management tools for preliminary facade screening and inspection documentation, providing supporting information for subsequent engineering review and maintenance planning. Full article

Timely computation of vegetation indices from remote sensing imagery would assist in sustainable agriculture monitoring through quick analysis of vegetation state and surface water. However, high-resolution multispectral imagery can be computationally expensive to process, especially on embedded systems. In this research work, we evaluate heterogeneous approaches that aim to enhance the computation of the Normalized Difference Vegetation Index (NDVI) and the Normalized Difference Water Index (NDWI) through sequential C++, OpenMP, CUDA and OpenCL on desktop and embedded CPU–GPU platforms. We compare multicore and GPU-based computations while also exploring optimizations of the OpenCL kernels with respect to memory management, loop unrolling and work-group settings. OpenMP increased the image-processing throughput from 211.24 to 500.85 images/s on the desktop platform and from 55.30 to 145.11 images/s on the Odroid XU4. These results correspond to speedups of 2.37× and 2.62×, respectively. These results demonstrate that heterogeneous processing can accelerate vegetation index calculation and provide a computational basis for timely, locally available agricultural monitoring. This capability may support precision agriculture applications, including vegetation stress assessment and water management decisions. Nevertheless, the experiments used offline multispectral images; energy consumption, energy per image, water savings, and onboard UAV performance were not measured. Full article

Mountainous forest-fire surveillance with multiple UAVs is constrained by rugged terrain, time-varying winds, and temperature-dependent battery derating, which jointly affect endurance and route feasibility. Most existing patrol models simplify these effects through planar routing and constant energy-consumption assumptions. This study develops an energy-aware mixed-integer linear programming model that integrates terrain-induced climbing, period-dependent wind conditions, and battery derating into multi-UAV mission scheduling. Two valid inequalities—a symmetry-breaking cut and a fleet-size lower-bound cut—are embedded in a branch-and-cut framework to improve exact solution efficiency. The proposed method is applicable to year-round patrol planning under season-dependent meteorological conditions. In the Jinyun Mountain, Chongqing, case study, it is evaluated under three representative seasonal scenarios (summer, autumn, and winter), with particular emphasis on the summer pre-fire period as the primary high-risk operating scenario. Within the same computational time limit, the proposed approach reduces total flight distance by approximately 36–49% and energy consumption by approximately 61–74% relative to large neighborhood search and ant colony optimization, while using fewer UAVs to complete the same patrol tasks. These results demonstrate that explicitly coupling environmental and battery constraints can improve the operational efficiency and energy feasibility of multi-UAV wildfire surveillance, and can provide practical decision support for daily wildfire-prevention patrol planning in mountainous environments. Full article

Climate-related and environmental hazards affect cultural heritage sites in markedly different inland, coastal, lacustrine and underwater settings, creating documentation requirements that cannot be addressed by a single sensing approach. This study presents the multi-sensor geometric documentation of eight cultural heritage sites. Unmanned aerial vehicle (UAV) photogrammetry was applied to six inland and coastal sites, while underwater photogrammetry, unmanned surface vehicles (USVs), acoustic sounding and a prototype green-wavelength flash LiDAR were used at three shallow-water sites. The campaigns produced orthomosaics, elevation models, dense point clouds, textured meshes, bathymetric maps and underwater LiDAR point clouds at scales appropriate to the conservation problem of each site. The resulting products document exposed architectural remains, excavation areas, cliffs and unstable slopes, lake-margin changes, submerged masonry, wooden structures and lakebed morphology. Their main contribution is the establishment of spatially explicit, site-specific baselines that provide measurable geometric and visual evidence for condition assessment, future repeat-survey comparisons and the spatial integration of environmental, archaeological and conservation information. The study demonstrates the operational and information complementarity of optical, acoustic and active ranging approaches, which address different documentation scales, environmental constraints and heritage targets, and provide distinct spatial evidence that can serve as potential inputs to subsequent digital twin and decision support applications. Full article

Methods for object localization and classification in images acquired from unmanned aerial vehicles (UAVs) quickly develop and find new applications. Pre-trained convolutional neural networks (CNNs) play the key role in solving these tasks. However, there are many factors that degrade the quality of acquired images and make the performance of methods intended for object detection and classification worse. One such factor is lossy compression of acquired images or video data widely used to pass them from on-board sensors and devices of preliminary data processing to on-land centers that perform further data processing for retrieval of valuable information. Both CNNs applied for localization and classification, and lossy compression techniques used to reduce the transferred data size have an impact on final results. To study this impact, we analyze the performance of several modifications of YOLO (You Only Look Once) CNNs applied to color images compressed by JPEG, which continues to be one of the basic compression tools. The quality factor is varied within wide limits to detect the situation when distortions due to lossy compression start to become too large and have a considerable negative effect on the localization and classification of objects of different types and sizes. Analysis is carried out using several traditional criteria, including Intersection over Union, F1, and mAP metrics, as well as some others. Dependence of localization and classification characteristics on the object size is performed. The datasets VisDrone and TAI are employed in training and verification. Full article

Unmanned aerial vehicles (UAVs) have become essential platforms for remote sensing in challenging environments such as high-dynamic-range (HDR) scenes and low-texture areas. However, conventional frame-based visual–inertial simultaneous localization and mapping (SLAM) systems often suffer from motion blur and overexposure during high-speed UAV flight, leading to state estimation failure. To address numerical instability in marginalization, weak scene adaptability, and insufficient outlier suppression in event-based stereo visual–inertial SLAM systems for aerial applications, we propose HESVI, a hybrid marginalization and adaptive heterogeneous kernel state estimation method for UAV remote sensing. Our method first establishes a focal-length-driven cross-modal inverse depth consistency constraint to couple image and event inverse depths, providing high-quality priors for optimization. Such lightweight prior generation is designed with the limited onboard computing resources of UAV platforms in mind. A hybrid marginalization strategy is then introduced, employing block-parallel tall–skinny QR (TSQR) acceleration based on Householder reflections alongside dynamic Tikhonov regularization and first-estimates Jacobian (FEJ) linearization to balance computational efficiency and numerical stability. Furthermore, an adaptive heterogeneous Cauchy kernel maps differentiated thresholds to image and event features according to their average effective tracking lengths, enabling dynamic outlier suppression. Experiments on the VECtor, MVSEC, and HKU datasets demonstrate that HESVI achieves the best absolute trajectory error (ATE) on the vast majority of the evaluated sequences, with average ATE reductions of 47.2%, 41.2%, and 26.8% over PL-EVIO, ESIO, and ESVIO, where each average is computed only over the sequences on which the corresponding baseline runs successfully. The method also exhibits excellent performance in complex remote-sensing scenarios and generalization tests. HESVI effectively enhances the numerical stability, scene adaptability, and localization accuracy of event-based stereo visual–inertial SLAM systems in challenging UAV remote-sensing environments. Full article

Vibration damper detection in unmanned aerial vehicle (UAV)-based transmission line inspection presents distinctive task-specific challenges: the targets are not only small and weakly textured, but also characterized by slender structures. Their effective identification therefore depends on the preservation of local contour cues and the appropriate organization of deep contextual responses. To address the limitations of conventional lightweight detectors in structural feature representation, cross-scale semantic consistency, and bounding-box localization, this paper proposes CARE-Net (Cascaded Attention and Refinement Enhanced Network), a compact detection framework for vibration damper detection. CARE-Net adopts an asymmetric design consisting of front-end structural enhancement and back-end contextual refinement. Specifically, the Cascaded Residual Attention Block (CRAB) is deployed in the backbone to strengthen the representation of slender contours and local structural features of vibration damper targets. The Dynamic Context Refinement Network (DCRN) is introduced at the backbone–neck transition to improve the contextual organization of deep features and the quality of cross-scale feature fusion. Meanwhile, an Adaptive Focal Complete IoU Loss (AF-CIoU) is proposed to optimize bounding-box regression for difficult samples without altering the inference architecture. A UAV-based vibration damper dataset covering three condition categories, namely normal, rusted, and dilapidated, is constructed in this study. Experimental results show that CARE-Net achieves an mAP^@0.5 of 0.951 and an mAP^@0.5:0.95 of 0.628 with 2.44 M parameters and 6.2 GFLOPs. Further configuration experiments indicate that, compared with repeatedly introducing attention enhancement into high-level features, stage-specific feature modeling is better suited to the slender small-object detection task investigated in this study. The proposed method provides a solution for intelligent vibration damper inspection of transmission lines that balances detection accuracy, model compactness, and potential for terminal-side application. Full article

The 3D geometric documentation of cultural heritage monuments requires spatial datasets that are accurate, complete and suitable for conservation, monitoring, visualization and heritage management. However, complex geometries, occlusions, limited accessibility, vegetation and other field-acquisition constraints often prevent a single surveying technique from providing a complete and metrically reliable representation. In this context, photogrammetry and simultaneous localization and mapping (SLAM)-based mapping provide complementary capabilities, with each method offering advantages and limitations regarding metric accuracy, spatial coverage, detail representation, acquisition flexibility and operational efficiency. This work develops, applies and evaluates a reproducible end-to-end workflow for the metric 3D documentation of complex cultural heritage monuments through multi-sensor integration. The proposed approach combines the metric robustness and visual richness of photogrammetric reconstruction with the rapid acquisition and spatial coverage enabled by SLAM-based mapping, while producing reusable datasets for conservation planning, comparative studies, education and broader heritage applications. The workflow integrates unmanned aerial vehicle (UAV) and close-range photogrammetry, SLAM-based mapping and geodetic control within a common reference system and is demonstrated through the documentation of a historic monastery. Both datasets showed centimetre-level agreement with geodetic observations, while photogrammetry yielded fuller exterior coverage and higher-quality texture, and SLAM enabled rapid interior coverage. The CH-PhotoSLAM3D dataset is released to support reproducibility and further research. Full article

Monitoring forest regeneration is vital for sustainable management, yet traditional ground surveys and early aerial imagery methods face cost, labor, and resolution limitations. Unmanned Aerial Vehicles (UAVs) offer cost-effective, flexible platforms for acquiring ultra-high-resolution data to address these gaps. This paper reviews recent research on UAV applications in forest regeneration surveys (FRS), tracing the evolution from field-based surveys and conventional aerial approaches to current UAV practices, and synthesizing developments in data acquisition, processing workflows, and analysis. It contrasts established Canopy Height Model (CHM) and point-cloud approaches with the growing use of deep learning, particularly Convolutional Neural Networks (CNNs) for seedling detection, crown delineation, density, height estimation, and species classification. Particular attention is given to accuracy assessment, examining sampling design, reference data, prediction-to-reference matching, and evaluation metrics, and highlighting the disconnect between traditional map-validation principles and standard deep-learning metrics that often neglect background classes. The case study applying Mask R-CNN to red pine seedlings in an Adirondack Park plantation achieved stand-level recall of 70.3% and precision of 98.7%, while plot-level DL detections represented only 36.8% of the field-observed seedling count. These results demonstrate the potential of DL for reliably identifying visible red pine seedlings while highlighting its limitations for complete regeneration inventories, particularly when seedlings have small crown sizes. Full article

Vision-based unmanned aerial vehicle (UAV) detection has become increasingly important since it is a key technology in aerial collision avoidance systems. However, the detection of small UAVs is still unsatisfactory in practical applications. To address this problem, this paper proposes EWH-YOLO, a novel deep convolutional neural network-based method for small UAV detection. First, an efficient feature extraction network is designed to exact the multi-level features of small UAVs while reducing the network parameters and computational complexity. Second, a weighted bidirectional feature fusion network is proposed to enhance the low-level and high-level features in the output feature maps. Third, a hybrid bounding box regression loss is introduced to evaluate the difference between the predicted bounding box and the ground-truth bounding box during training and improve the detection accuracy. Finally, a new dataset is created on the basis of considering small UAVs to verify the detection performance. Compared with the state-of-the-art methods, the proposed method achieves higher detection accuracy with lower model complexity. The experimental results demonstrate that the proposed detector significantly improves the detection performance of small UAVs. Full article

While there are evident advantages to deploying UAVs for structural inspection applications within reinforced-concrete structures where human access may be hazardous or restricted, UAV deployment is still inhibited within structures due to a lack of GPS, payload and power limitations, signal attenuation, limited transmission opportunities, and low bandwidth for communication. Within the confines of the present project, a low-cost two-layer six-rotor architecture was developed; a single Pixhawk PX4 manages stabilized flight, while a Raspberry Pi 4 manages mission-level operations. The complete prototype integrated a relative-pose/nearby-object estimator, hybrid 433 MHz and Wi-Fi/MQTT communications, avionics-side energy management, mission continuity via SQLite, and human-in-the-loop fail-safe support. Prototype experiments revealed reduced horizontal drift compared to both tested comparison configurations, continued operation under constrained communication conditions, and a 62.0% reduction in avionics-side power consumption, excluding propulsion. This was followed by a complementary PX4–Gazebo evaluation probing horizontal and vertical proximity responses, six-sector LiDAR processing, and stale-data watchdog functionality. Across 45 repeated simulation runs and 3000 retained sector-level observations, no simulated collisions occurred during the horizontal-approach, vertical-proximity, or watchdog tests. No false sector assignments were observed, and the overall mean absolute error was 0.0285 m. From this limited demonstration, the architecture appears satisfactory at the prototype and simulated-subsystem levels. Further physical testing will be needed prior to operational deployment. Full article