Search Results (571)

Professional UAV thermal imaging systems are widely used for inspection, environmental monitoring, search and rescue, agriculture, and technical diagnostics. However, their cost limits their use in education, preliminary field screening, rapid prototyping, and low-resource applications. This study evaluates a minimum-cost indirect UAV thermal sensing workflow based on a DJI Mini 4K consumer drone, a lightweight Servo King9000 smartphone, and a UTi260M smartphone-connected infrared thermal camera. In the proposed configuration, the smartphone displayed and recorded the thermal stream, while the onboard RGB camera of the UAV recorded the smartphone-displayed infrared video during flight. The aim was not to develop a radiometric UAV thermal imaging platform, but to determine whether such a low-cost configuration can provide qualitative presence/absence indication of clear thermal hotspots and to identify its operational limits. The system was experimentally assessed under no-payload and payload conditions, daylight and nighttime illumination, and several low-altitude operating heights. Additional motor-region thermal observations were performed using a UTi260T handheld thermal camera under loaded and unloaded operating conditions. The complete UAV–payload configuration had a measured mass of approximately 340 g, corresponding to an effective added payload of 91 g and a payload-to-UAV mass ratio of 36.5%. Payload operation reduced near-ground flight endurance from approximately 25 min to 14 min 40 s. The maximum observed motor-region temperature increased from 24.9 °C under unloaded operation to 42.0 °C under loaded operation, while motor thermal asymmetry increased from 4.8 °C to 7.6 °C. Nighttime and low-glare operation improved the readability of the smartphone-displayed thermal stream, with the most practical usability observed at approximately 10–20 m. The results show that the proposed workflow is feasible only for short-range qualitative thermal screening and clear hotspot presence/absence indication. The UAV-recorded video should not be interpreted as direct thermal data, but as an RGB recording of a smartphone display showing thermal information. Therefore, the workflow is not suitable for quantitative temperature measurement, radiometric thermal mapping, or accurate thermal shape delineation. The main operational limits are payload mass, suspended-load oscillation, display readability, reduced endurance, motor-region thermal loading, sensitivity to payload alignment, and the absence of raw radiometric data. Direct UTi260M smartphone-recorded thermal frames were additionally used for pixel-size-assisted qualitative verification of practical reference thermal targets, including a human-sized target and a vehicle-sized target, at selected low-altitude operating heights. Full article

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. Full article

Individual tree species classification is essential for detailed forest inventories, ecosystem monitoring, and biodiversity assessment. While UAV-acquired RGB and multispectral (MS) imagery have advanced tree species mapping, most studies focus on a single sensor type. In practice, UAV platforms carry diverse sensors with varying spatial resolutions, spectral bands, radiometric responses, and noise characteristics, introducing domain shifts that limit model generalization across datasets. To overcome these challenges, we propose a supervised cross-sensor transfer learning approach, leveraging a DenseNet-121 model pretrained on high-resolution UAV RGB imagery to improve classification on lower-resolution multispectral imagery with limited labelled data. The adapted model achieved 75% overall accuracy and a macro-F1 score of 0.706, significantly improving over models trained from scratch. Its performance was further evaluated on downsampled UAV MS imagery simulating conventional airborne multispectral photographs, demonstrating robustness and practical applicability for regional-scale forest inventories. This study highlights cross-domain transfer learning as a pathway toward sensor-independent, efficient, and operationally scalable tree species classification. Full article

The canopy closure stage is a critical phase of rice (Oryza sativa L.) development that influences canopy structure and final grain yield. Accurate and continuous monitoring of canopy closure dynamics is therefore essential for variety screening and cultivation optimization. This study combines unmanned aerial vehicle (UAV) remote sensing technology with deep learning-based semantic segmentation to establish an efficient framework for quantifying rice canopy closure dynamics. UAV RGB images were acquired for 198 hybrid rice varieties during early growth stages and used to build a canopy segmentation dataset. Three semantic segmentation models, i.e., DeepLabv3+, U-Net, and PSPNet, were systematically evaluated. Results show that DeepLabv3+ performed the best and enabled precise extraction of rice canopy features, obtaining a mean intersection over union (mIoU) of 0.86. Based on the extracted canopy coverage, the Gompertz model was utilized to characterize temporal canopy closure trajectories for all varieties, achieving an average R² of 0.978. Subsequently, five key dynamic indicators were derived, including canopy closure limit value (K), initial growth coefficient (a), growth rate coefficient (b), maximum instantaneous growth rate (MGR), and days to maximum growth rate (Tm). K-means clustering analysis was performed on these indicators to categorize all rice varieties into three clusters, disclosing pronounced differences in early-stage canopy development characteristics. Correlation analysis further demonstrated that canopy closure dynamics were closely associated with grain yield. Overall, while acknowledging the limitations of a single-season and single-site dataset, this study provides a scalable and objective framework for quantifying rice canopy closure dynamics, offering valuable support for variety selection, cultivation optimization, and high-yield rice production. Full article

Unmanned aerial vehicle (UAV) red–green–blue–infrared (RGB-IR) object detection is important for traffic monitoring, security surveillance, and urban management, but remains challenging because aerial targets are often small, densely distributed, and affected by complex backgrounds. In addition, RGB and infrared (IR) modalities contribute unequally under different imaging conditions, making simple feature concatenation or indiscriminate middle-layer fusion insufficient for stable cross-modal utilization. To address this problem, this paper proposes Selective Interaction Mechanism and Prefiltering Complementary Spatial Refinement (SIM-PCSR), a key-layer complementary enhancement method for UAV RGB-IR small-object detection. The proposed method decomposes cross-modal modeling into two stages. SIMAdapter first performs selective interaction on the small-object-sensitive P3 layer before fusion, suppressing redundant responses and enhancing potentially complementary modal evidence. PCSR then refines the fused representation through prefiltering, modal selection, and local window residual refinement, injecting reliable complementary information into the key-layer fused feature in a controlled manner. Experiments on the DroneVehicle dataset show that SIM-PCSR achieves 85.323 mean average precision (mAP)₅₀ and 63.572 mAP_50:95, improving the Fixed Middle Fusion baseline by 0.523 and 0.751 percentage points, respectively. These gains correspond to relative improvements of 0.62% and 1.20% over the baseline. Module ablation, position ablation, repeated-seed evaluation, category-wise analysis, scale-wise analysis, and qualitative visualization jointly demonstrate that explicit selection and organization of cross-modal information can improve UAV RGB-IR small-object detection under modality imbalance and background interference. Full article

Drone-to-drone (D2D) detection is a critical yet underexplored task in low-altitude intelligent perception, where UAV targets are often small, weakly textured, motion-affected, and disturbed by complex backgrounds and environmental changes. Existing cross-domain detection datasets mainly focus on ground objects or single-factor shifts, making them insufficient for evaluating D2D detection under coupled real-world variations. To address this gap, we present TriCross-D2D, an RGB air-to-air UAV detection dataset and benchmark with three explicit domain shifts: scene, viewpoint, and weather. Built from real flight videos and controlled synthetic fog, TriCross-D2D contains 13 RGB video sequences, 23,403 raw frames, 7045 benchmark images, and 9771 annotated UAV instances. It provides a fixed split of 4045 Source_train images, 2000 Target_train images, and 1000 Target_val images, supporting both unsupervised domain adaptation (UDA) and semi-supervised domain adaptation (SSDA). The dataset is dominated by small objects, with extremely tiny, tiny, and small targets accounting for 73.8% of all instances. Benchmark results show that existing cross-domain detectors still perform limitedly on TriCross-D2D, especially under stricter localization and recall metrics. Single-factor analysis further reveals that the coupled scene–viewpoint–weather protocol is more challenging than isolated shifts, with viewpoint variation producing a particularly strong domain gap. As an exploratory enhanced baseline, SCOPE-DA-RTDETR improves DA-RTDETR from 28.63/13.12/22.39 to 29.94/13.71/23.40 in ${\mathrm{AP}}_{50}/{\mathrm{AP}}_{50–95}/\mathrm{AR}$, showing consistent but modest gains. These findings demonstrate that TriCross-D2D provides a challenging and discriminative benchmark for cross-domain D2D small-object detection. Full article

Early wildfire monitoring in power transmission corridors requires reliable detection of weak fire and smoke cues under complex field conditions and strict edge-computing constraints. To address these issues, this paper proposes an edge-deployable RGB–thermal framework based on visible and thermal infrared (TIR) imaging for unmanned aerial vehicle (UAV)-based corridor monitoring, including a spatial detector, YOLO-MMSC, and a temporal-enhanced version, YOLO-MMSC-T. The study also establishes a self-collected corridor-oriented RGB–thermal (RGB–T) dataset to complement public wildfire data. Unlike existing RGB–thermal wildfire datasets that mainly focus on forest or wildland fire scenes, the proposed dataset is specifically organized for complex-background power transmission-corridor monitoring, including continuous UAV sequences, nighttime conditions, smoke/vegetation occlusion, long-range small targets, and hard-negative interference. To the best of our knowledge, this is the first self-collected RGB–thermal wildfire dataset designed for this specific application scenario. The framework integrates a mobile inverted bottleneck convolution (MBConv) lightweight backbone, a Shallow Detail Fusion Module (SDFM) for shallow cross-modal alignment and denoising, a Content-Guided Attention (CGA) module for adaptive fusion, and normalized Wasserstein distance (NWD)-based box regression for long-range small-target localization. Experiments on public and self-collected datasets show that YOLO-MMSC achieves 94.6% mAP@0.5, 95.0% precision, and 93.9% recall while running at 60 FPS on Jetson Orin NX. With temporal fine-tuning, YOLO-MMSC-T reaches a continuous detection rate (CDR) of 95.6% with a jitter index of $2.8\times {10}^{-3}$. Field experiments using a DJI Matrice 4T further indicate a practical operating altitude of 120–180 m. These results support lightweight RGB–thermal remote sensing for real-time wildfire monitoring in complex transmission-corridor environments. Full article

Accurate tree counting from remote sensing data is essential for forest inventory, biomass estimation, carbon accounting, and ecological monitoring. However, existing approaches predominantly rely on airborne RGB imagery and often struggle in complex forest scenes where neighboring crowns exhibit highly similar textures and colors and where overlapping crown boundaries become ambiguous. To address this limitation, the LiDAR-derived Canopy Height Model (CHM) is introduced as a complementary modality that provides explicit cues on canopy height variation and vertical structure to support RGB-based analysis. Building on this, we propose BCAR-Net, a broker-guided RGB and depth (RGB-D) multimodal framework that couples bidirectional cross-modal interaction, adaptive tri-branch fusion, and auxiliary reconstruction within a two-stage optimization scheme. Specifically, a bidirectional cross-attention U-Net generates an intermediate broker RGB-D representation from paired RGB images and depth maps through symmetric bidirectional cross-attention between the two modalities and direction-aware gating. The original RGB image, depth map, and broker representation are then jointly encoded by three weight-sharing branches and adaptively aggregated by a spatial fusion gate for density-map regression. To regularize the fused latent feature, a multi-scale cross-attention reconstruction decoder provides auxiliary RGB and depth reconstruction supervision by querying multi-scale BCA-UNet encoder features through 2D cross-attention, and a reconstruction-oriented first stage replaces externally generated fused-image supervision, yielding a task-consistent optimization scheme. Experiments on the NEONTreeEvaluation benchmark show that BCAR-Net consistently outperforms single-modality settings and direct RGB-D concatenation multimodal baseline. Additional experiments on a public UAV RGB–LiDAR dataset provide a small-scale supplementary evaluation under a different acquisition setting, where BCAR-Net achieves modest but consistent improvements over RGB-only and depth-only baselines. These results demonstrate that the proposed framework offers an effective but computationally cautious solution for tree counting in complex forest environments. Full article

Accurate crack width measurement is important for structural condition assessment, but image-based methods are sensitive to oblique viewing angles and varying imaging distances. To address this challenge, this study proposes a method and evaluation framework for crack width measurement under non-orthogonal imaging conditions using RGB-D sensors. The proposed method integrates plane fitting and homography-based geometric rectification to transform imaged cracks into standard orthogonal viewpoints. It then applies dynamic masking and hybrid global–local binarization to the rectified image to improve measurement accuracy and robustness. Finally, this study develops an evaluation framework for comparing the proposed and baseline methods under different viewing angles and imaging distances. The framework establishes correspondences between physical locations along the same crack across RGB-D images captured under different imaging conditions, enabling quantitative analysis of performance variations. Experiments on two cracks in concrete buildings show that the proposed method outperforms the baseline method without geometric rectification, reducing the fitted surface error by 19.3–52.3% while maintaining a validity rate above 99%. The results indicate that incorporating surface geometry offers a practical pathway for quantitative crack assessment in close-range image-based inspection using handheld or UAV-mounted RGB-D cameras. Full article

High-throughput and non-destructive phenotyping approaches are increasingly needed to support precision agriculture and plant breeding. This study evaluates the use of unmanned aerial vehicle (UAV) multispectral imagery combined with machine learning to estimate anthocyanin content in red lettuce genotypes under field conditions. High-resolution RGB and multispectral images were acquired using a low-cost UAV platform, and vegetation indices sensitive to pigment variation were extracted at the plot scale. Ridge regression, decision tree, and random forest models were trained using 80% of the dataset and validated with the remaining 20%. Random forest achieved the highest performance for anthocyanin estimation, with coefficients of determination reaching R² = 0.84 and lower prediction errors than linear approaches. Overall, the results demonstrate that UAV-based multispectral sensing integrated with machine learning provides a robust, scalable, and cost-effective solution for non-destructive pigment phenotyping, with direct applications in biofortification-oriented breeding and precision agriculture. Full article

Accurate registration of UAV-derived thermal infrared orthophotos and RGB orthophotos is essential for multi-sensor geospatial analysis, but it remains challenging because thermal imagery generally has lower spatial resolution, weaker texture, and less distinct structural information than RGB imagery. This study comparatively evaluated five histogram equalization methods—histogram equalization (HE), contrast-limited adaptive histogram equalization (CLAHE), brightness-preserving bi-histogram equalization (BBHE), dualistic sub-image histogram equalization (DSIHE), and minimum mean brightness error bi-histogram equalization (MMBEBHE)—for improving AKAZE-based registration of land surface temperature (LST) orthophotos to reference RGB orthophotos. High-accuracy RGB orthophotos generated using GNSS-surveyed ground control points were used as the geometric reference. Thermal data were acquired twice at each of two study sites with contrasting surface characteristics and processed into LST orthophotos. Each histogram equalization method was applied to the LST orthophotos, after which keypoints and descriptors were extracted using AKAZE, tentative correspondences were established, outliers were removed using RANSAC, and an affine transformation was estimated from the inlier correspondences. Here, an inlier denotes a tentative match that remained geometrically consistent after RANSAC-based outlier rejection. The estimated transformation was then applied to the source LST raster to preserve radiometric values in the final corrected product. Performance was assessed using the number of detected keypoints, tentative matches, RANSAC-verified inliers, matching efficiency, reproducibility, and exploratory statistical analysis. Among the five methods, BBHE consistently produced the highest number of inliers and the best matching efficiency at both study sites, while also showing the lowest variability between repeated acquisitions. These results indicate that brightness-preserving histogram equalization is particularly effective for thermal–RGB orthophoto registration and can improve the reliability of UAV-derived thermal mapping products for geomatics applications. Full article

Destructive methods for monitoring stress responses remain a bottleneck in precision agriculture. This study presents a non-destructive methodological framework evaluating drought responses in 30 Opuntia ficus-indica plants over four months under five irrigation levels. Cladode traits (color, weight, and thickness) were measured alongside RGB imagery from a UAV and hyperspectral imaging (400–1000 nm). Partial least squares regression (PLSR) models showed high capability to model proline (R² = 0.91), chlorophyll a (R² = 0.97), and total chlorophyll (R² = 0.97) within the experimental dataset. Crucially, these models reflected continuous spectral–physiological variation across the irrigation gradient rather than discrete treatment separation, with key spectral regions identified at 530–600 nm and 550–750 nm. UAV-derived RGB imagery enabled the estimation of plant area and biomass (R² = 0.88). Under extreme drought, cladode thickness decreased by approximately 41%, accompanied by reduced biomass and increased soluble solids (°Brix). While no statistically significant differences were observed among irrigation treatments for biochemical variables, limiting treatment discrimination based on discrete classification, the hyperspectral data successfully captured the underlying continuous physiological variation. Consequently, this work demonstrates the methodological viability of integrating UAV structural phenotyping and hyperspectral analysis as a continuous monitoring tool rather than a rigid classification system. These findings provide a methodological baseline that highlights the need for continuous sensing in CAM plants, though further validation with independent datasets remains essential for wider operational application. Full article

The early detection and spatial characterization of crop damage are critical for improving decision-making in precision agriculture, particularly in regions where traditional monitoring methods are limited in scalability and objectivity. This study presents an integrated information processing framework that couples UAV-based image acquisition, instance segmentation, slicing-aided inference of large orthomosaics, and georeferenced spatial analysis into a single reproducible pipeline for the detection and mapping of crop damage. The framework is applied to maize cultivated under traditional milpa systems in Yucatán, Mexico, a region characterized by intercropping, irregular plant spacing, and complex backgrounds rarely represented in mainstream agricultural deep learning benchmarks. High-resolution RGB images were systematically acquired over maize fields in Yucatán, Mexico, and curated into specialized datasets representing parcels, individual plants, and damaged vegetation. Instance segmentation models based on the YOLOv11 architecture were trained and evaluated to extract visual information related to crop condition, while the Slicing-Aided Hyper Inference (SAHI) method was integrated to enable efficient processing of large orthomosaic images. The proposed framework achieved high performance in detecting maize plants, with a precision of 92.9% and an mAP50 of 94.2%, and demonstrated reliable identification of damage patterns associated with Spodoptera frugiperda, reaching a precision of 79.2% and an mAP50 of 71.7%. The resulting georeferenced outputs provide spatially explicit information that supports quantitative analysis of crop health and damage distribution. The results indicate that the proposed framework constitutes a scalable and reproducible approach for UAV-based visual information extraction, with potential applicability to broader agricultural monitoring and data-driven decision support systems. Full article

Forest ecosystems in northeastern Japan are characterized by natural mixed forests, where sustainable management has always been limited because of their difficult accessibility. The aims of this study are first, to assess mixed forest composition, and second, to train Mask R-CNN models with these data in order to detect and segment trees in a 19-ha mixed forest composed mainly of beech (Fagus crenata), oak (Quercus crispula), magnolia (Magnolia obovata) and larch (Larix kaempferi). The Mask R-CNN model was applied in two experimental scenarios: a single multi-class model and species-specific models. RGB images consisted of four orthomosaics (August, September, October 2024 and October 2025), which yielded 1725, 359, 129 and 525 samples of each tree species, respectively. The Unmanned Aerial Vehicle (UAV)-QField validation method improved the classification accuracy of the annotations and made it possible to map each tree species distribution and understand the composition of mixed forests along an elevation gradient. The multi-class model demonstrated an overall precision of 0.59, a recall of 0.53, and an F1-score of 0.56. The detection performance for individual tree species was similar for both models. Based on these results, the multi-class model is more suitable because it decreases the possibility of misclassification of tree species. Full article

Maize is a globally important staple crop, and automated monitoring of germination and seedling emergence is essential for precision agriculture, enabling timely reseeding and reducing potential yield loss. To address this need, we propose Seedling-DETR, a transformer-based model for the real-time detection of emerged and missing maize seedlings using multispectral UAV imagery in an end-to-end manner. First, we construct a multispectral UAV dataset and introduce a dedicated annotation strategy in which missing seedlings were labeled individually rather than inferred indirectly. Then, we modify the feature fusion module of RT-DETR and develop a hybrid-scale feature fusion module to obtain richer and more expressive feature representations for missing seedling detection and improve the precision of missing seedling detection. Finally, we propose a channel fusion module to incorporate multispectral images into our model without requiring a dedicated multispectral backbone or additional pretraining, thereby improving model adaptability. The results show that, under a random train–test split (8:2), when using RGB images as input, our Seedling-DETR achieves a mean average precision (mAP) of 83.1% at an IoU threshold of 0.5, outperforming YOLOv11x and RT-DETR by 2.5% and 1.1%, respectively. The proposed method achieves an AP of 69.3% at an IoU threshold of 0.5 for missing seedling detection, which increases to 71.7% when multispectral inputs are incorporated. Similar performance trends are observed on an independent validation set collected on a different date. Although the model introduces moderate computational overhead (282 GFLOPs for RGB input and 418 GFLOPs for multispectral configuration, with 84.0 M and 85.1 M parameters, respectively), it can maintain efficient detection performance suitable for actual agricultural field deployment. The method is further validated at the field scale using orthomosaic-based analysis. Overall, this study provides an effective and scalable framework for the detection of emerged and missing maize seedlings under complex field conditions. The proposed framework supports accurate reseeding decisions, and contributes to automated maize production. Full article