Search Results (559)

This study combined virtual reality technology with drone aerial imagery of Smangus, a remote Atayal tribe situated 1500 m above sea level in Hsinchu County, Taiwan, to develop a virtual drone system. This study aims to investigate the learning effectiveness and operational experience associated with the application of the virtual drone system for exploring tribal natural landscapes and enhancing junior high school students’ learning of Indigenous settlement site selection. A quasi-experimental design was conducted with two seventh-grade classes from a junior high school in Hsinchu County, Taiwan. The experimental group (n = 43) engaged with the virtual drone system to perform settlement site selection tasks, while the control group (n = 42) learned using traditional materials such as PowerPoint slides and maps. The intervention consisted of two instructional sessions, with data collected via achievement tests, questionnaires, and open-ended feedback. The results indicated that students in the experimental group significantly outperformed the control group in learning outcomes. Positive responses were also observed in learning motivation, cognitive load, and system satisfaction. Students reported that the virtual drone system improved students’ understanding of terrain and enhanced their skills in selecting appropriate sites while increasing their interest and motivation in learning. Moreover, the course incorporated the Atayal people’s migration history and field interview data, enriching its cultural authenticity and contextual relevance. Full article

In Latin America, high seismic activity drives countries to develop disaster risk reduction policies based on seismic risk studies. This work demonstrates the feasibility of creating a seismic exposure and vulnerability database using remotely sensed data. In Nejapa, El Salvador, a drone flight and 360° photo capture were conducted to generate a 3D model of the city. Buildings were identified, characterised, and assigned a vulnerability model. This database was used to estimate seismic risk for a simulated Mw 6.7 earthquake on the Guaycume fault near the city. Results show that 71% of buildings would suffer complete damage and 68% of the population would be homeless, with losses exceeding USD 15 million. Findings were shared with relevant institutions in El Salvador through a dashboard. The country is currently collecting the same type of data used in the present study to update its cadastre and census. This is an opportunity to replicate this pilot experience in many other cities across the country and to provide open data access, positioning El Salvador at the forefront of civil protection in the Latin American region. Full article

In low-altitude multi-object tracking (MOT), challenges such as frequent inter-object occlusion and complex non-linear motion disrupt the appearance of individual targets and the continuity of their trajectories, leading to frequent tracking failures. We posit that the relatively stable spatio-temporal relationships within object groups (e.g., pedestrians and vehicles) offer powerful contextual cues to resolve such ambiguities. We present NOWA-MOT (Neighbors Know Who We Are), a novel tracking-by-detection framework designed to systematically exploit this principle through a multi-stage association process. We make three primary contributions. First, we introduce a Low-Confidence Occlusion Recovery (LOR) module that dynamically adjusts detection scores by integrating IoU, a novel Recovery IoU (RIoU) metric, and location similarity to surrounding objects, enabling occluded targets to participate in high-priority matching. Second, for initial data association, we propose a Graph Cross-Attention (GCA) mechanism. In this module, separate graphs are constructed for detections and trajectories, and a cross-attention architecture is employed to propagate rich contextual information between them, yielding highly discriminative feature representations for robust matching. Third, to resolve the remaining ambiguities, we design a cascaded Matched Neighbor Guidance (MNG) module, which uniquely leverages the reliably matched pairs from the first stage as contextual anchors. Through MNG, star-shaped topological features are built for unmatched objects relative to their stable neighbors, enabling accurate association even when intrinsic features are weak. Our comprehensive experimental evaluation on the VisDrone2019 and UAVDT datasets confirms the superiority of our approach, achieving state-of-the-art HOTA scores of 51.34% and 62.69%, respectively, and drastically reducing identity switches compared to previous methods. Full article

In recent years, drone-based wildfire detection technology has advanced rapidly, yet existing methods still encounter numerous challenges. For instance, high background complexity leads to frequent false positives and false negatives in models, which struggle to accurately identify both small-scale fire points and large-scale wildfires simultaneously. Furthermore, the complex model architecture and substantial parameter count hinder lightweight deployment requirements for drone platforms. To this end, this paper presents a lightweight drone-based wildfire detection model, DFE-YOLO. This model utilizes dynamic frequency domain enhancement technology to resolve the aforementioned challenges. Specifically, this study enhances small object detection capabilities through a four-tier detection mechanism; improves feature representation and robustness against interference by incorporating a Dynamic Frequency Domain Enhancement Module (DFDEM) and a Target Feature Enhancement Module (C2f_CBAM); and significantly reduces parameter count via a multi-scale sparse sampling module (MS3) to address resource constraints on drones. Experimental results demonstrate that DFE-YOLO achieves mAP50 scores of 88.4% and 88.0% on the Multiple lighting levels and Multiple wildfire objects Synthetic Forest Wildfire Dataset (M4SFWD) and Fire-detection datasets, respectively, whilst reducing parameters by 23.1%. Concurrently, mAP50-95 reaches 50.6% and 63.7%. Comprehensive results demonstrate that DFE-YOLO surpasses existing mainstream detection models in both accuracy and efficiency, providing a reliable solution for wildfire monitoring via unmanned aerial vehicles. Full article

Norway spruce (Picea abies (L.) H. Karst.) and Scots pine (Pinus sylvestris L.) are among the most valuable tree species in the Lithuanian forests. Pure stands, which comprise approximately one-quarter of Lithuania’s forest area, provide an important framework for studying tree responses to thinning and susceptibility to species-specific diseases and damage. This study investigated stem health and quality in two experimental Scots pine stands (32 and 39 years old) and four experimental Norway spruce stands (36–43 years old) to assess the influence of the initial stand density and thinning intensity. Each stand consisted of five plots with different initial densities and was subjected to varying thinning regimes from stand establishment. Tree locations were mapped using the pseudolite-based positioning system TerraHärp, and local tree density was calculated. Stem health and damage were assessed using ICP-Forests methodology. Our results showed that across initial densities of 1000–4400 trees ha⁻¹, tree dimensions (diameter and height) were similar, regardless of thinning intensity. The highest levels of stem damage and competition-induced mortality occurred in the densest, unthinned stands, with deer browsing and scraping from fallen trees being the most common damage agents. In contrast, thinned stands exhibited a higher incidence of stem rot (Heterobasidion annosum (Fr.) Bref.), particularly for Norway spruce. Finally, stand density alone did not consistently explain the patterns of tree mortality in either the pine or spruce stands. These findings suggest that cultivating Scots pine and Norway spruce at lower initial densities with minimal thinning may reduce the damage and losses caused by fungal infection. Finally, novel techniques, such as the pseudolite-based positioning system for geolocating trees and drone imaging for assessing tree health, have proven valuable in facilitating field surveys. Full article

Current drone navigation systems primarily rely on Global Navigation Satellite Systems (GNSSs), but their signals are susceptible to interference, spoofing, or suppression in complex environments, leading to degraded positioning performance or even failure. To enhance the positioning accuracy and robustness of low-altitude drones in satellite-denied environments, this paper investigates an absolute visual localization solution. This method achieves precise localization by matching real-time images with reference images that have absolute position information. To address the issue of insufficient feature generalization capability due to the complex and variable nature of ground scenes, a visual-based image retrieval algorithm is proposed, which utilizes a fusion of shallow spatial features and deep semantic features, combined with generalized average pooling to enhance feature representation capabilities. To tackle the registration errors caused by differences in perspective and scale between images, an image registration algorithm based on cyclic consistency matching is designed, incorporating a reprojection error loss function, a multi-scale feature fusion mechanism, and a structural reparameterization strategy to improve matching accuracy and inference efficiency. Based on the above methods, a hierarchical absolute visual localization system is constructed, achieving coarse localization through image retrieval and fine localization through image registration, while also integrating IMU prior correction and a sliding window update strategy to mitigate the effects of scale and rotation differences. The system is implemented on the ROS platform and experimentally validated in a real-world environment. The results show that the localization success rates for the h, s, v, and w trajectories are 95.02%, 64.50%, 64.84%, and 91.09%, respectively. Compared to similar algorithms, it demonstrates higher accuracy and better adaptability to complex scenarios. These results indicate that the proposed technology can achieve high-precision and robust absolute visual localization without the need for initial conditions, highlighting its potential for application in GNSS-denied environments. Full article

This paper explores the physical layer security performance of collaborative drone fleets enabled by visible light communication (VLC) in a multi-eavesdropper scenario, where multiple drones leverage VLC to serve terrestrial users. To strengthen system security, we formulate a sum worst-case secrecy rate maximization problem. To address the non-convex optimization challenge of this problem, we develop two innovative Q-learning-based position decision algorithms (Q-PDA and Q-PDA-lite) with a dynamic reward mechanism, allowing drones to adaptively optimize their positions. Additionally, we propose an enhanced Tabu Search-based grouping algorithm (TS-GA) to establish the suboptimal user equipment (UE)–drone association by balancing candidate solution exploration and tabu constraint exploitation. Simulation results demonstrate that the proposed Q-PDA and Q-PDA-lite achieve worst-case secrecy rates significantly exceeding those of Random-PDA and K-means-PDA. While Q-PDA-lite exhibits 2% lower performance than Q-PDA, it offers reduced complexity. Additionally, the proposed TS-GA achieves a worst-case secrecy rate that substantially outperforms random grouping, UE-channel-gain-based grouping, and channel-gain-based grouping. Collectively, the hybrid approach integrating Q-PDA and TS-GA achieves 10% near-global optimality with guaranteed convergence, while preserving computational efficiency. Furthermore, this hybrid approach outperforms other combinations in terms of security metrics. Full article

Multi-drone operations face significant efficiency challenges when launch pad locations are predetermined without optimization, leading to suboptimal route configurations and increased travel distances. This research addresses launch pad positioning as a continuous planar location-routing problem (PLRP), developing a genetic algorithm framework integrated with multiple Traveling Salesman Problem (mTSP) solvers to optimize launch pad coordinates within operational areas. The methodology was evaluated through extensive experimentation involving over 17 million test executions across varying problem complexities and compared against brute-force optimization, Particle Swarm Optimization (PSO), and simulated annealing (SA) approaches. The results demonstrate that the genetic algorithm achieves 97–100% solution accuracy relative to exhaustive search methods while reducing computational requirements by four orders of magnitude, requiring an average of 527 iterations compared to 30,000 for PSO and 1000 for SA. Smart initialization strategies and adaptive termination criteria provide additional performance enhancements, reducing computational effort by 94% while maintaining 98.8% solution quality. Statistical validation confirms systematic improvements across all tested scenarios. This research establishes a validated methodological framework for continuous launch pad optimization in UAV operations, providing practical insights for real-world applications where both solution quality and computational efficiency are critical operational factors while acknowledging the simplified energy model limitations that warrant future research into more complex operational dynamics. Full article

The rapid evolution of UAV technology has increased the demand for lightweight airborne perception systems. This study introduces ADG-YOLO, an optimized model for real-time target detection and ranging on UAV platforms. Building on YOLOv11n, we integrate C3Ghost modules for efficient feature fusion and ADown layers for detail-preserving downsampling, reducing the model’s parameters to 1.77 M and computation to 5.7 GFLOPs. The Extended Kalman Filter (EKF) tracking improves positional stability in dynamic environments. Monocular ranging is achieved using similarity triangle theory with known target widths. Evaluations on a custom dataset, consisting of 5343 images from three drone types in complex environments, show that ADG-YOLO achieves 98.4% mAP_0.5 and 85.2% mAP_0.5:0.95 at 27 FPS when deployed on Lubancat4 edge devices. Distance measurement tests indicate an average error of 4.18% in the 0.5–5 m range for the DJI NEO model, and an average error of 2.40% in the 2–50 m range for the DJI 3TD model. These results suggest that the proposed model provides a practical trade-off between detection accuracy and computational efficiency for resource-constrained UAV applications. Full article

To address the challenges of small target detection in UAV aerial images—such as difficulty in feature extraction, complex background interference, high miss rates, and stringent real-time requirements—this paper proposes an innovative model series named SRTSOD-YOLO, based on YOLO11. The backbone network incorporates a Multi-scale Feature Complementary Aggregation Module (MFCAM), designed to mitigate the loss of small target information as network depth increases. By integrating channel and spatial attention mechanisms with multi-scale convolutional feature extraction, MFCAM effectively locates small objects in the image. Furthermore, we introduce a novel neck architecture termed Gated Activation Convolutional Fusion Pyramid Network (GAC-FPN). This module enhances multi-scale feature fusion by emphasizing salient features while suppressing irrelevant background information. GAC-FPN employs three key strategies: adding a detection head with a small receptive field while removing the original largest one, leveraging large-scale features more effectively, and incorporating gated activation convolutional modules. To tackle the issue of positive-negative sample imbalance, we replace the conventional binary cross-entropy loss with an adaptive threshold focal loss in the detection head, accelerating network convergence. Additionally, to accommodate diverse application scenarios, we develop multiple versions of SRTSOD-YOLO by adjusting the width and depth of the network modules: a nano version (SRTSOD-YOLO-n), small (SRTSOD-YOLO-s), medium (SRTSOD-YOLO-m), and large (SRTSOD-YOLO-l). Experimental results on the VisDrone2019 and UAVDT datasets demonstrate that SRTSOD-YOLO-n improves the mAP@0.5 by 3.1% and 1.2% compared to YOLO11n, while SRTSOD-YOLO-l achieves gains of 7.9% and 3.3% over YOLO11l, respectively. Compared to other state-of-the-art methods, SRTSOD-YOLO-l attains the highest detection accuracy while maintaining real-time performance, underscoring the superiority of the proposed approach. Full article

Precise estimate of antenna location is essential for high-quality three-dimensional (3D) radar imaging, especially under sparse sampling schemes. In scenarios involving synchronized scanning and rotational motion, small deviations in the radar’s transmitting position can lead to significant phase errors, thereby degrading image fidelity or even causing image failure. To address this challenge, we propose a novel trajectory estimation method based on microwave three-point ranging. The method utilizes three fixed microwave-reflective calibration spheres positioned outside the imaging scene. By measuring the one-dimensional radial distances between the radar and each of the three spheres, and geometrically constructing three intersecting spheres in space, the radar’s spatial position can be uniquely determined at each sampling moment. This external reference-based localization scheme significantly reduces positioning errors without requiring precise synchronization control between scanning and rotation. Furthermore, the proposed approach enhances the robustness and flexibility of sparse sampling strategies in near-field radar imaging. Beyond ground-based setups, the method also holds promise for drone-borne 3D imaging applications, enabling accurate localization of onboard radar systems during flight. Simulation results and error analysis demonstrate that the proposed method improves trajectory accuracy and supports high-fidelity 3D reconstruction under non-ideal sampling conditions. Full article

The waste dump from the abandoned La Soterraña mine, a former mercury extraction site, contains high concentrations of mercury (Hg) and arsenic (As), which pose a significant environmental risk due to direct exposure to the environment. Given the site’s topography and slope, surface runoff has been identified as the primary mechanism for the dispersal of these toxic elements into nearby watercourses. This study quantifies the amount of Hg and As released into fluvial systems through surface runoff from the waste dump. Historical topographic data, Airborne Laser Exploration Survey public data from the National Plan for Aerial Orthophotographs (1st PNOA-LiDAR) of the Spanish Ministry of Transport, Mobility and Urban Agenda, and high-precision photogrammetric drone surveys were utilized, with centimeter-level accuracy achieved using airborne GNSS RTK positioning systems on the drone. The methodology yields reliable results when comparing surfaces generated from topographic data collected with consistent methodologies and standards. Analysis indicates an environmental release exceeding 1000 kg of mercury (Hg) and 12,000 kg of arsenic (As) between 2019 and 2023, based on high spatial resolution data (GSD = 8 cm). These findings highlight a sustained temporal contribution of chemical contaminants, which imposes serious environmental and biological health risks due to persistent exposure to toxic elements. Full article

With the development of drone technology in recent years, many studies have discussed how to leverage drones equipped with sensors and cameras to conduct inspections under bridges. To address positioning challenges caused by the lack of GPS signals under the bridges, triangulation methods with on-site pre-installed Ultra-Wideband (UWB) sensors were used extensively to determine drone locations. However, the practical hurdles of deploying anchors under bridges are often overlooked, including variable terrain and potential electromagnetic interference from deploying a large number of UWB sensors. This study introduces a handover mechanism to address long-distance positioning challenges and an enhanced two-stage algorithm to enhance its suitability for bridge terrain with higher stability. By integrating these concepts, a practical bridge inspection system was devised, and realistic under-bridge experiments were conducted to validate the method’s efficacy in real-world settings. Full article

McMahon Hall, an iconic Brutalist dormitory at the University of Washington, has become the site of an interdisciplinary experiment in cultural memory and machine-assisted storytelling. This article presents a method that combines remote sensing with AI-generated voices to produce a polyvocal narrative of architecture through the perspective of the building itself, its material (concrete), an architect, a journalist, and a bird. Drone photogrammetry and generated 3D models were combined with generative AI (text, image, and voice) to reconstruct the site digitally and imaginatively (AI-driven speculative narratives). Through speculative storytelling, the article and the project explore how cultural memory and perception of built heritage can be augmented by machines, offering plural perspectives that challenge singular historical narratives. The Introduction situates the work at the intersection of digital heritage documentation, AI storytelling, epistemology in machine learning, and spatial computing, emphasizing the perception of heritage through different actors. The Theoretical Framework draws on literature in photogrammetry for heritage preservation, polyvocal narrative, and knowledge frameworks of AI. The Materials and Methods detail the workflow: capturing McMahon Hall via UAV photogrammetry, producing a 3D model, and generating character-driven narratives with large language models and voice synthesis. The resulting multi-voiced narrative and its thematic insights are described. In the Discussion, the implications of this approach for architectural heritage interpretation are considered, including its capacity to amplify diverse voices and the risks of bias or hyperreality in AI-generated narratives. The study argues that this polyvocal, machine-augmented storytelling expands the toolkit of remote sensing and digital heritage by not only documenting the tangible form of the built environment but also speculating on its intangible cultural memory. The Conclusions reflect on how merging spatial computing techniques with AI narratives can support new modes of engagement with architecture, positioning this work as a building block toward richer human-machine co-created heritage experiences. Full article

In this paper, we propose a novel method to reduce non-Gaussian errors in measurements using sampling-based distribution estimation. Although non-Gaussian errors are often treated as statistical deviations, they can frequently arise in practical unmanned aerial systems that depend on global navigation satellite systems (GNSS), where position measurements are degraded by multipath effects. However, nonlinear or robust filters have shown limited effectiveness in correcting such errors, particularly when they appear as persistent biases in measurements over time. In such cases, adaptive techniques have often demonstrated greater effectiveness. The proposed method estimates the distribution of observed measurements using a sampling-based approach and derives a reformed measurement from this distribution. By incorporating this reformed measurement into the filter update, the proposed approach achieves lower error levels than traditional adaptive filters. To validate the effectiveness of the method, Kalman filter simulations are conducted for drone GNSS/INS navigation. The results show that the proposed method outperforms conventional non-Gaussian filters in handling measurement bias caused by non-Gaussian errors. Furthermore, it achieves nearly twice the estimation accuracy compared to adaptive approaches. These findings confirm the robustness of the proposed technique in scenarios where measurement accuracy temporarily deteriorates before recovering. Full article