Unmanned Aerial Vehicles Research Articles

Machine Learning (AutoML)-Driven Wheat Yield Prediction for European Varieties: Enhanced Accuracy Using Multispectral UAV Data

Accurate and timely wheat yield prediction is valuable globally for enhancing agricultural planning, optimizing resource use, and supporting trade strategies. Study addresses the need for precision in yield estimation by applying machine-learning (ML) regression models to high-resolution Unmanned Aerial Vehicle (UAV) multispectral (MS) and Red-Green-Blue (RGB) imagery. Research analyzes five European wheat cultivars across 400 experimental plots created by combining 20 nitrogen, phosphorus, and potassium (NPK) fertilizer treatments. Yield variations from 1.41 to 6.42 t/ha strengthen model robustness with diverse data. The ML approach is automated using PyCaret, which optimized and evaluated 25 regression models based on 65 vegetation indices and yield data, resulting in 66 feature variables across 400 observations. The dataset, split into training (70%) and testing sets (30%), was used to predict yields at three growth stages: 9 May, 20 May, and 6 June 2022. Key models achieved high accuracy, with the Support Vector Regression (SVR) model reaching R2 = 0.95 on 9 May and R2 = 0.91 on 6 June, and the Multi-Layer Perceptron (MLP) Regressor attaining R2 = 0.94 on 20 May. The findings underscore the effectiveness of precisely measured MS indices and a rigorous experimental approach in achieving high-accuracy yield predictions. This study demonstrates how a precise experimental setup, large-scale field data, and AutoML can harness UAV and machine learning’s potential to enhance wheat yield predictions. The main limitations of this study lie in its focus on experimental fields under specific conditions; future research could explore adaptability to diverse environments and wheat varieties for broader applicability.

Accurate Evaluation of Urban Mangrove Forest Health Considering Stand Structure Indicators Based on UAVs

Stand structural configuration dictates ecosystem functional performance. Mangrove ecosystems, located in ecologically sensitive coastal ecotones, require efficient acquisition of stand structure parameters and health assessments based on these parameters for practical applications. Effective assessment of mangrove ecosystem health, crucial for their functional performance in ecologically sensitive coastal ecotones, relies on efficient acquisition of stand structure parameters. This study developed a UAV (Unmanned Aerial Vehicle)-based framework for mangrove health evaluation integrating stand structure parameters, utilizing UAV visible-light imagery, field plot surveys, and computer vision techniques, and applied it to the assessment of a national nature reserve. We obtained the following results: (1) A deep neural network, combining UAV visible-light data with tree height constraints, achieved 88.29% overall accuracy in simultaneously identifying six dominant mangrove species; (2) Stand structure parameters were derived based on individual tree extraction results in seedling zones along forest edges (with canopy individual tree segmentation accuracy ≥ 78.57%), and a stand health evaluation model was constructed; (3) Health assessment revealed that the core zone exhibited significantly superior stand health compared to non-core zones. This method demonstrates high efficiency, significantly reducing the time and effort for monitoring, and offers robust support for future mangrove forest health assessments and adaptive conservation strategies.

Design and Structural Evaluation of a 6-DOF Robotic Arm

This study lays out the design of a robotic arm with six degrees of freedom created to automate the battery swapping operation of unmanned aerial vehicles (UAVs) employed in agriculture. Because UAVs have limited flight capability, the batteries need to be changed frequently, affecting negatively the continuity and efficiency of operations. The robotic arm designed was 3D-modeled in SolidWorks and put through transient structural analyses in Ansys to determine its behavior in actual conditions. A specially designed cycloidal gear system was incorporated in every joint of the robotic arm for high torque output, low backlash, and compactness. The robot's modular design makes it suitable for different UAV models and mission scenarios. The selection of materials was done considering strength, weight, and manufacturability, and the outer housing is made of aluminum A356, while internal parts are made of 1.8550 steel alloy. The gripper at the end-effector is pneumatically actuated and capable of handling battery modules of up to 12 kg safely. Inverse kinematics solutions, dynamic simulations, and transient load analyses validated the structural stiffness, low displacement, and high precision of the robotic arm. The findings confirm the effectiveness of the system for field applications, and it is a feasible and novel approach to automated UAV battery replacement in agriculture.

Research on Real-Time Monitoring Algorithm for Edge-Based Aviation Meteorological Disasters Integrating Multi-source Data

Unmanned aerial vehicles (UAVs) can take advantage of cutting-edge precision provided by artificial intelligence (AI), particularly when it comes to deep learning (DL) focused methods. This could improve the UAVs’ remote sensing features for a variety of evacuation and disaster prevention possibilities. To lessen the impacts of disasters on the surroundings and human beings as quickly as possible, UAVs with camera lenses can operate in distant and challenging-to-access disaster-affected areas, examine the images, and sound alerts when different disasters, like collapsed structures, floods, or fires, occur. The implementation of these kinds of neural systems in numerous contexts requiring low-delay boundaries on deduction and mission-critical choices in real-time is hindered by the hefty processing needs introduced by the incorporation of DL. For disaster mitigation and monitoring uses, this paper presents a unique atrous-residual fusion network (ARFN) for effective aerial image categorization from drones. To train the proposed method, multiple sensed samples are acquired and processed on the Python platform. The performance of the proposed method is analyzed and compared with other existing methods. The investigation is conducted using a Python tool, and performance metrics such as F1-score (94.9%), recall (95.6%), accuracy (92.7%) and precision (94.6%) are utilized to evaluate the research’s effectiveness. To handle multiple resolution attributes on low-power systems effectively and increase efficiency, the design suggested makes use of atrous convolutions. From the experimental results, it can be concluded that the proposed approach attains the greatest performance of the existing methods in disaster management.

Efficient Hotspot Detection in Solar Panels via Computer Vision and Machine Learning

Solar power generation is rapidly emerging within renewable energy due to its cost-effectiveness and ease of deployment. However, improper inspection and maintenance lead to significant damage from unnoticed solar hotspots. Even with inspections, factors like shadows, dust, and shading cause localized heat, mimicking hotspot behavior. This study emphasizes interpretability and efficiency, identifying key predictive features through feature-level and What-if Analysis. It evaluates model training and inference times to assess effectiveness in resource-limited environments, aiming to balance accuracy, generalization, and efficiency. Using Unmanned Aerial Vehicle (UAV)-acquired thermal images from five datasets, the study compares five Machine Learning (ML) models and five Deep Learning (DL) models. Explainable AI (XAI) techniques guide the analysis, with a particular focus on MPEG (Moving Picture Experts Group)-7 features for hotspot discrimination, supported by statistical validation. Medium Gaussian SVM achieved the best trade-off, with 99.3% accuracy and 18 s inference time. Feature analysis revealed blue chrominance as a strong early indicator of hotspot detection. Statistical validation across datasets confirmed the discriminative strength of MPEG-7 features. This study revisits the assumption that DL models are inherently superior, presenting an interpretable alternative for hotspot detection; highlighting the potential impact of domain mismatch. Model-level insight shows that both absolute and relative temperature variations are important in solar panel inspections. The relative decrease in “blueness” provides a crucial early indication of faults, especially in low-contrast thermal images where distinguishing normal warm areas from actual hotspot is difficult. Feature-level insight highlights how subtle changes in color composition, particularly reductions in blue components, serve as early indicators of developing anomalies.

Research on situation awareness strategy for drone swarm defense

Low-altitude safety is key to the sustainable development of the low-altitude economy. Drone swarms pose greater risks than individual unmanned aerial vehicles due to their scale and coordination. This paper proposes a situation awareness strategy for the defense of drone swarm. Multi-scale drone target detection is achieved through an anchor-free structure, drone swarm formation recognition is realized by Graph Neural Networks, and the situation of drone swarm is calculated by constructing macroscopic quantitative descriptors. It breaks through the feature extraction and fusion algorithm for multi-scale drones, graph neural networks for intra-layer and inter-layer feature extraction, and macroscopic quantitative descriptors based on divergence and curl to construct scale-invariant and rotation-invariant features. It achieves the detection of whether it is a drone swarm, the identification of which drone swarm it is, and the calculation of the degree of the drone swarm, providing a basis for the classification and graded handling of drone swarm, and effectively promoting the modernization of the low-altitude safety governance capacity.

Thermal Infrared UAV Applications for Spatially Explicit Wildlife Occupancy Modeling

Assessing the impact of community-based conservation programs on wildlife biodiversity remains a significant challenge. This pilot study was designed to develop and demonstrate a scalable, spatially explicit workflow using thermal infrared (TIR) imagery and unmanned aerial vehicles (UAVs) for non-invasive biodiversity monitoring. Conducted in a 2-hectare grassland area in Chitwan, Nepal, the study applied TIR-based grid sampling and multi-species occupancy models with thin-plate splines to evaluate how species detection and richness might vary between (1) morning and evening UAV flights, and (2) the Chitwan National Park and Kumroj Community Forest. While the small sample area inherently limits ecological inference, the aim was to test and demonstrate data collection and modeling protocols that could be scaled to larger landscapes with sufficient replication, and not to produce generalizable ecological findings from a small dataset. The pilot study results revealed higher species detection during morning flights, which allowed us to refine our data collection. Additionally, models accounting for spatial autocorrelation using thin plate splines suggested that community-based conservation programs effectively balanced ecosystem service extraction with biodiversity conservation, maintaining richness levels comparable to the national park. Models without splines indicated significantly higher species richness within the national park. This study demonstrates the potential for spatially explicit methods for monitoring grassland mammals using TIR UAV as indicators of anthropogenic impacts and conservation effectiveness. Further data collection over larger spatial and temporal scales is essential to capture the occupancy more generally for species with larger home ranges, as well as any effects of rainfall, flooding, and seasonal variability on biodiversity in alluvial grasslands.

SR-YOLO: Spatial-to-Depth Enhanced Multi-Scale Attention Network for Small Target Detection in UAV Aerial Imagery

The detection of aerial imagery captured by Unmanned Aerial Vehicles (UAVs) is widely employed across various domains, including engineering construction, traffic regulation, and precision agriculture. However, aerial images are typically characterized by numerous small targets, significant occlusion issues, and densely clustered targets, rendering traditional detection algorithms largely ineffective for such imagery. This work proposes a small target detection algorithm, SR-YOLO. It is specifically tailored to address these challenges in UAV-captured aerial images. First, the Space-to-Depth layer and Receptive Field Attention Convolution are combined, and the SR-Conv module is designed to replace the Conv module within the original backbone network. This hybrid module extracts more fine-grained information about small target features by converting image spatial information into depth information and the attention of the network to targets of different scales. Second, a small target detection layer and a bidirectional feature pyramid network mechanism are introduced to enhance the neck network, thereby strengthening the feature extraction and fusion capabilities for small targets. Finally, the model’s detection performance for small targets is improved by utilizing the Normalized Wasserstein Distance loss function to optimize the Complete Intersection over Union loss function. Empirical results demonstrate that the SR-YOLO algorithm significantly enhances the precision of small target detection in UAV aerial images. Ablation experiments and comparative experiments are conducted on the VisDrone2019 and RSOD datasets. Compared to the baseline algorithm YOLOv8s, our SR-YOLO algorithm has improved mAP@0.5 by 6.3% and 3.5% and mAP@0.5:0.95 by 3.8% and 2.3% on the datasets VisDrone2019 and RSOD, respectively. It also achieves superior detection results compared to other mainstream target detection methods.

Estimating Biomass in Eucalyptus globulus and Pinus pinaster Forests Using UAV-Based LiDAR in Central and Northern Portugal

Accurate biomass estimation is important for forest management and climate change mitigation. This study evaluates the potential of using LiDAR (Light Detection and Ranging) data, acquired through Unmanned Aerial Vehicles (UAVs), for estimating above-ground and total biomass in Eucalyptus globulus and Pinus pinaster stands in central and northern Portugal. The acquired LiDAR point clouds were processed to extract structural metrics such as canopy height, crown area, canopy density, and volume. A multistep variable selection procedure was applied to reduce collinearity and select the most informative predictors. Multiple linear regression (MLR) models were developed and validated using field inventory data. Random Forest (RF) models were also tested for E. globulus, enabling a comparative evaluation between parametric and machine learning regression models. The results show that the 25th height percentile, canopy cover density at two meters, and height variance demonstrated an accurate biomass estimation for E. globulus, with coefficients of determination (R2) varying between 0.86 for MLR and 0.90 for RF. Although RF demonstrated a similar predictive performance, MLR presented advantages in terms of interpretability and computational efficiency. For P. pinaster, only MLR was applied due to the limited number of field data, yet R2 exceeded 0.80. Although absolute errors were higher for Pinus pinaster due to greater biomass variability, relative performance remained consistent across species. The results demonstrate the feasibility and efficiency of UAV LiDAR point cloud data for stand-level biomass estimation, providing simple and effective models for biomass estimation in these two species.

Dynamic Bearing–Angle for Vision-Based UAV Target Motion Analysis

The Bearing–Angle algorithm effectively improves the observability of vision-based motion estimation for moving targets by combining the dimensional information of target detection frames. However, the robustness of this algorithm will be significantly reduced when the observation error increases due to sudden changes in the target motion state. To address this shortcoming, this paper proposes a visual target motion estimation algorithm called the Dynamic Bearing–Angle, which aims to improve the accuracy and robustness of target motion analysis in dynamic scenarios such as unmanned aerial vehicle (UAV). The algorithm innovatively introduces a dual robustness mechanism of dynamic noise intensity adaptation and outlier suppression based on M-estimation. By adjusting the noise covariance matrix in real time and assigning low weights to the outlier observations using the Huber weight function, the Dynamic Bearing–Angle algorithm is able to effectively cope with non-Gaussian noise and sudden target maneuvers. We validate the performance of the proposed algorithm with numerical simulations and real sensor data, and the results show that the Dynamic Bearing–Angle maintains good robustness and accuracy under different noise intensities.

UAV-OVD: Open-Vocabulary Object Detection in UAV Imagery via Multi-Level Text-Guided Decoding

Object detection in drone-captured imagery has attracted significant attention due to its wide range of real-world applications, including surveillance, disaster response, and environmental monitoring. Although the majority of existing methods are developed under closed-set assumptions, and some recent studies have begun to explore open-vocabulary or open-world detection, their application to UAV imagery remains limited and underexplored. In this paper, we address this limitation by exploring the relationship between images and textual semantics to extend object detection in UAV imagery to an open-vocabulary setting. We propose a novel and efficient detector named Unmanned Aerial Vehicle Open-Vocabulary Detector (UAV-OVD), specifically designed for drone-captured scenes. To facilitate open-vocabulary object detection, we propose improvements from three complementary perspectives. First, at the training level, we design a region–text contrastive loss to replace conventional classification loss, allowing the model to align visual regions with textual descriptions beyond fixed category sets. Structurally, building on this, we introduce a multi-level text-guided fusion decoder that integrates visual features across multiple spatial scales under language guidance, thereby improving overall detection performance and enhancing the representation and perception of small objects. Finally, from the data perspective, we enrich the original dataset with synonym-augmented category labels, enabling more flexible and semantically expressive supervision. Experiments conducted on two widely used benchmark datasets demonstrate that our approach achieves significant improvements in both mean mAP and Recall. For instance, for Zero-Shot Detection on xView, UAV-OVD achieves 9.9 mAP and 67.3 Recall, 1.1 and 25.6 higher than that of YOLO-World. In terms of speed, UAV-OVD achieves 53.8 FPS, nearly twice as fast as YOLO-World and five times faster than DetrReg, demonstrating its strong potential for real-time open-vocabulary detection in UAV imagery.

Comparative Analysis of Energy Efficiency and Position Stability of Sub-250 g Quadcopter and Bicopter with Similar Mass Under Varying Conditions

This paper investigates the energy efficiency and positional stability of two types of ultralight unmanned aerial vehicles (UAVs)—bicopter and quadcopter—both with mass below 250 g, under varying flight conditions. The study is motivated by increasing interest in low-weight drones due to their regulatory flexibility and application potential in constrained environments. A comparative methodology was adopted, involving the construction of both UAV types using identical components where possible, including motors, sensors, and power supply, differing only in propulsion configuration. Experimental tests were conducted in wind-free and wind-induced environments to assess power consumption and stability. The data were collected through onboard blackbox logging, and positional deviation was tracked via video analysis. Results show that while the quadcopter consistently demonstrated lower energy consumption (by 6–22%) and higher positional stability, the bicopter offered advantages in simplicity of frame design and reduced component count. However, the bicopter required extensive manual tuning of PID parameters due to the inherent instability introduced by servo-based control. The findings highlight the potential of bicopters in constrained applications, though they emphasize the need for precise control strategies and high-performance servos. The study fills a gap in empirical analysis of energy consumption in lightweight bicopter UAVs.

LADOS: Aerial Imagery Dataset for Oil Spill Detection, Classification, and Localization Using Semantic Segmentation

Oil spills on the water surface pose a significant environmental hazard, underscoring the critical need for developing Artificial Intelligence (AI) detection methods. Utilizing Unmanned Aerial Vehicles (UAVs) can significantly improve the efficiency of oil spill detection at early stages, reducing environmental damage; however, there is a lack of training datasets in the domain. In this paper, LADOS is introduced, an aeriaL imAgery Dataset for Oil Spill detection, classification, and localization by incorporating both liquid and solid classes of low-altitude images. LADOS comprises 3388 images annotated at the pixel level across six distinct classes, including the background. In addition to including a general oil class describing various oil spill appearances, LADOS provides a detailed categorization by including emulsions and sheens. Detailed examination of both instance and semantic segmentation approaches is illustrated to validate the dataset’s performance and significance to the domain. The results on the test set demonstrate an overall performance exceeding 66% mean Intersection over Union (mIoU), with specific classes such as oil and emulsion to surpass 74% of IoU part of the experiments.

Integrated Maintainability in the Structural and System Design of Fixed-Wing UAVs

This paper introduces an overall design approach to a fixed-wing unmanned aerial vehicle (UAV) with an emphasis on the incorporation of maintainability into structural and system-level designs. Conceptual design was established using CAD software, following which aerodynamic performance was analyzed by external flow simulations based on the finite volume method. The outcome of the above analyses was used to finalize the external shape, resulting in detailed structural and layout designs. Mechanical behavior of critical structural components was analyzed through finite element analysis (FEA) to determine stress distribution, deformation, and vibration characteristics under simulated operational loading conditions. A distinctive feature of this research is the prioritization of maintainability as a primary design criterion from the outset. The inner arrangement of components; batteries, electric ducted fan motors, avionics, and control surfaces, was carefully planned to provide maintainability access for both personnel and tools. Sensor and control systems were also embedded with the aim of facilitating condition-based and predictive maintenance over UAV’s operational life. This involved choosing sensor points of attachment and providing access to areas expected to need regular inspection, adjustment, or replacement. The completed UAV prototype optimizes aerodynamics, structural strength, and maintainability. The design strategy not only satisfies performance and production demand but also helps to reduce maintenance downtime and lifecycle costs. This maintainability-driven design philosophy serves to increase the UAV's viability for actual world deployment in environments in which reliability, serviceability, and operational readiness are vital.

Palm Oil Health Monitoring Based on Vegetation Index through Unmanned Aerial Vehicles (UAV)/Drone Tools

The plantation sector has an urgency in technological renewal and development, especially in palm oil. Palm oil commodities are included in the staples needed by the Indonesian people for various products and as one of the sectors that generate the largest foreign exchange. Palm oil plantation management practices that have been carried out manually are less able to control plant health, especially during the pandemic which has caused severe impacts in Indonesia. The purpose of this research is to showcase the implementation of Smart Agriculture through the utilization of Unmanned Aerial Vehicle (UAV)/Drone technology in determining palm oil health. This research uses spatial analysis to detect the level of palm oil health by utilizing aerial photographs and GIS software with the Visible Atmospherically Resistant Index (VARI) method which can detect differences in the wavelength of the spectrum reflected by the plant canopy. This research is located in Argomulyo Village, Sepaku District, Penajam Paser Utara Regency, East Kalimantan Province, with abundant palm oil plantations. It was found that out of 2,167 samples of palm oil plants in the studied plantation, 14% were very unhealthy, 34.61% were unhealthy, 36.54% were healthy, and 14.85% were very healthy.

A crack detection and quantification method using matched filter and photograph reconstruction

Crack detection is a critical task for bridge maintenance and management. While popular deep learning algorithms have shown promise, their reliance on large, high-quality training datasets, which are often unavailable in engineering practice, limits their applicability. By contrast, traditional digital image processing methods offer low computational costs and strong interpretability, making continued research in this area highly valuable. This study proposes an automatic crack detection and quantification approach based on digital image processing combined with unmanned aerial vehicle (UAV) flight parameters. First, the characteristics of the bridge images collected by the UAVs were thoroughly analyzed. An enhanced matched-filter algorithm was designed to achieve crack segmentation. Morphological methods were employed to extract the skeletons of the segmented cracks, enabling the calculation of actual crack lengths. Finally, a 3D model was constructed by integrating the detection results with the image-shooting parameters. This 3D model, annotated with detected cracks, provides an intuitive and comprehensive representation of bridge damage, facilitating informed decision making in maintenance planning and resource allocation. To verify the accuracy of the enhanced matched filter algorithm, it was compared with other digital image processing methods on public datasets, achieving average results of 97.9% for Pixel Accuracy (PA), 72.5% for the F1-score, and 58.1% for Intersection over Union (Iou) across three typical sub-datasets. Moreover, the proposed methodologies were successfully applied to an arch bridge with an error of only 2%, thereby demonstrating their applicability to real-world scenarios.

A homotopy estimation based temporal-spatial spectrum prediction for UAV communications with arbitrary flight paths

Due to the rapid growth of unmanned aerial vehicles (UAVs), their spectrum resources become scarce, leading to UAVs requiring spectrum prediction to share the spectrum with other users. However, contemporary prediction methods may have difficulty in predicting the spectrum states at the next location, because the UAVs cannot obtain the historical data in advance to train prediction models. This paper introduces a temporal-spatial spectrum prediction approach for arbitrary flight within a specific region. The main issue involves the estimation of historical data at the next location during flight, accomplished through the concept of homotopy theory (HT). First, the HT is extended from two objects to multiple objects. Then, the historical data is estimated by homotopy mapping, which is derived by the boundary conditions of the HT and the physical meanings of the model parameters. Finally, the spectrum is predicted by the hidden Markov model (HMM) using the HT estimated data, referring to the multiple objects HT-HMM (MOHT-HMM) based prediction method. The main innovation is to use the HT to estimate the historical data at the next location, avoiding the non-stationarity and correlation issues of the spectra. Experimental results using real measured civil aviation data show the efficacy of the MOHT-HMM in accurately predicting UAV spectrum during arbitrary flights within a preset area.

Fusion of Satellite and UAV Imagery for Crop Monitoring

Abstract. Crop monitoring is crucial for precision agriculture, providing insights for optimizing yield and managing resources effectively. This study explores the fusion of Unmanned Aerial Vehicle (UAV) and Sentinel-2 (S2) satellite imagery for monitoring the crop by analyzing vegetation indices and canopy height information from the temporal dataset. Brovey Transform (BT) and Principal Component Analysis (PCA) fusion techniques are used to fuse the UAV and satellite images, aiming to leverage the high spatial resolution of UAV imagery with the broader spectral range of S2 data. Five key vegetation indices, including NDVI, GNDVI, SAVI, EVI, and LAI, were calculated from UAV, S2, and fused imagery in various temporal dates. Canopy height was derived from UAV data, and statistical analyses, including coefficient of determination (R2), Pearson correlation coefficient, and Root Mean Square Error (RMSE), were performed to assess relationships between canopy height and vegetation indices across the fused images and UAV and S2 images. Results indicate that fused imagery significantly enhances crop health metrics' accuracy and spatial relevance, with high R2 values and strong correlations between vegetation indices of fused images and UAV images, suggesting enhanced predictive power in monitoring crop health. Our findings highlight the advantages of fusing UAV and S2 imagery for comprehensive crop condition assessment, demonstrating that fused images provide a robust tool for monitoring crop vigor and stress levels. This approach offers valuable support for timely, data-driven decisions in crop management practices.

Utilization of Unmanned Aerial Vehicle for Monitoring and Surveillance of Aquaculture Farms: A Proposed Framework

Abstract— In this paper, unmanned aerial vehicles (UAVs) for monitoring and surveillance within aquaculture farms situated in Iloilo, Philippines are investigated. The principal aim is to create a framework that can provide guidance on how to use UAVs for effective monitoring, data collection, and regulatory control within the aquaculture sector under the supervision of – Department of Agriculture in the Municipality of Barotac and Dumangas, Iloilo. The study examines the performance of UAVs in real-time surveillance mapping as well as geotagging and geo- referencing that would help in observing fish behavior and environmental settings. The research assesses the precision of UAVs in executing these assignments and compares various models based on their effectiveness in monitoring perimeters environmental data collecting and following changes in fish behavior. Findings reveal that UAVs especially those with high detection precision levels and advanced mapping capabilities may be good tools for improving management practices on aquaculture farms. The paper suggests some ways we may adopt UAV technology within aquaculture emphasizing its potential to minimize operational costs while at the same time increasing productivity. With the help of these findings, it is possible to develop initial guidelines for UAV use in the aquaculture industry that will serve as a basis for a regulatory framework ensuring the sustainable and efficient functioning of aquaculture farms. This framework seeks to support local government initiatives in enhancing farm monitoring, protecting the environment and managing resources by using advanced UAV technology.

Automatic Detection of Tiny Drainage Outlets and Ventilations on Flat Rooftops from Aerial Imagery

Abstract. Flat rooftops on residential and industrial buildings house critical drainage and ventilation systems, which play essential roles in channeling water away from structures and preventing moisture accumulation. These utilities are vital for maintaining the structural integrity of rooftops, safeguarding against water pooling and moisture buildup that could otherwise lead to damage or even collapse, particularly during extreme weather events. However, current inspection and maintenance practices for these systems are predominantly manual, making them time-consuming, labor-intensive, and sometimes hazardous. This paper presents an automated approach to detecting drainage outlets and ventilation systems on flat rooftops, using a custom-labeled dataset of highresolution aerial imagery. We evaluated two different object detection methods, with FCOS (Fully Convolutional One-Stage Object Detection) outperforming Faster R-CNN in identifying these small utilities. The outcomes pave the way for new applications, as detected utilities can act as sparse data points that trigger constraint-based reasoning processes for estimating hidden utility networks in as-built Building Information Modeling (BIM) contexts. Embedding these identified objects into GIS or BIM models represents an initial step towards coarse-to-fine visual recognition, enabling customized semantic mission planning for autonomous exploration and inspection using Unmanned Aerial Vehicles (UAVs). The labeled dataset used in this study is publicly available by following this link https://zenodo.org/records/14040571.