Search Results (1,402)

Accurate rice-lodging mapping from unmanned aerial vehicle (UAV) imagery supports post-disaster loss assessment, crop insurance, and precision field management. Existing deep-learning methods typically require dense pixel-level annotations, which are costly and time-consuming to produce. Moreover, RGB imagery alone often fails to distinguish lodged from healthy rice when their canopy colors and textures are similar. To address these challenges, we propose Geometry-Guided UniMatch (GUMatch), a semi-supervised multimodal segmentation framework that leverages registered RGB imagery and UAV-derived digital surface models (DSMs). Unlike conventional approaches that treat DSMs as uniformly fused auxiliary channels, GUMatch uses them as reliability-aware geometric priors, incorporating height and boundary evidence to guide lodging segmentation. Our framework integrates three key components. First, Adaptive Geometric Prompting (AGP) injects DSM-based prompt features into the decoder based on local geometric reliability and RGB–DSM compatibility. Second, Geometry-Calibrated Pseudo-Label Learning (GPL) down-weights uncertain pseudo-label supervision within teacher-identified boundary-risk regions. Third, Boundary-Aware Geometric Regularization (BGR) refines boundary localization exclusively where pseudo-labels and geometric evidence are jointly reliable. Experiments are conducted on a three-parcel UAV rice-lodging collection. The main semi-supervised benchmark trains and selects models on the Huai’an parcel HA-P2 under labeled ratios of 10%, 20%, and 40%, and evaluates them on the held-out HA-P1 parcel. The external Wuxi parcel WX-P3 is reserved solely for direct cross-region testing. With an RN-101 backbone, GUMatch achieves 80.12%, 82.98%, and 85.06% mIoU under the three labeled ratios, consistently outperforming representative semi-supervised baselines including UniMatch V2 and RSProtoSemiSeg. With a DINOv2-B backbone, GUMatch further reaches 82.64%, 84.95%, and 86.72% mIoU. On WX-P3 under the 40% labeled setting with models trained on HA-P2, GUMatch with DINOv2-B achieves 68.34% mIoU, improving over UniMatch V2 by 5.39 points. These results demonstrate that reliability-aware geometric guidance enhances annotation-efficient UAV-based rice-lodging mapping. Full article

Permafrost degradation is intensifying differential settlement, structural deformation, and secondary periglacial hazards along linear infrastructure in cold regions, underscoring the need for monitoring approaches that integrate corridor-scale screening with fine-scale quantification. This study investigated highways, railways, transmission tower foundations, and buried pipelines in the permafrost region of Northeast China using multi-temporal UAV-borne LiDAR point clouds and synchronous visible-light imagery acquired by a DJI Matrice 300 unmanned aerial vehicle equipped with a DJI Zenmuse L1 sensor (DJI, Shenzhen, China). A synergistic optical–LiDAR framework was developed for distress identification and multidimensional quantification. The overall root mean square errors (RMSEs) at flight altitudes of 50 m and 100 m were 3.25 cm and 4.13 cm, respectively. By integrating texture and boundary information from synchronous visible-light imagery, elevation and volumetric metrics from LiDAR-derived digital elevation models (DEMs) and digital surface models (DSMs), and structural attitude parameters extracted from three-dimensional (3D) models, the framework enabled the parametric quantification of pavement cracking, differential shoulder settlement, railway embankment slump, transmission tower inclination, thaw settlement and ponding in pipeline trenches, and secondary icing. Snow-depth retrievals agreed well with field measurements (R² = 0.87, RMSE = 1.32 cm), indicating that UAV-LiDAR can extend monitoring into snow-covered periods. These findings provide a methodological basis for distress detection, screening of hazard-prone sections, and risk-informed operation and maintenance of linear infrastructure in permafrost regions. Full article

Multi-angle measurements provide essential data on the anisotropic reflectance properties of vegetation, enabling more robust retrievals of leaf area index (LAI), clumping index (CI), and canopy gap fraction compared to conventional single-view remote sensing. While Unmanned Aerial Vehicles (UAVs) offer unprecedented centimeter-level spatial resolution and flexible deployment, their application exposes a fundamental scale mismatch between ultra-high-resolution imagery and traditional bidirectional reflectance distribution function (BRDF) models. This review explicitly identifies that conventional 1D radiative transfer models (RTMs), which rely on the assumption of a statistically homogeneous canopy, suffer from severe scale-dependent biases, such as systematically underestimating hotspot reflectance (e.g., observed biases of 25% to 40% in 5-cm resolution UAV studies over specific vegetation canopies), and structural-optical confounding when directly applied to UAV data. At centimeter scales, macroscopic structural heterogeneity disrupts this homogeneity, necessitating the use of 3D RTMs that can explicitly simulate geometric occlusion and complex multiple scattering processes in highly heterogeneous environments. To bridge these theoretical and operational gaps, this review uniquely synthesizes UAV-specific multi-angle methodologies, systematically correlating canopy architectural types with optimal sensor configurations, flight strategies, and BRDF modeling frameworks. By evaluating recent advancements in multimodal data fusion, physics-informed machine learning, and physiological parameter retrieval, this review provides a comprehensive roadmap for decoupling structural and biochemical traits, highlighting how multi-angle directional signatures can substantially elevate classification accuracy, with specific experiments on spectrally similar crops and mixed tree species demonstrating improvements from roughly 40% to over 89%. Ultimately, it establishes practical, decision-oriented guidelines for overcoming transient illumination and co-registration errors, advancing high-fidelity quantitative monitoring and stress detection in complex forest ecosystems. Full article

Human-oriented smart cities and landscape systems increasingly depend on reusable, high-resolution geospatial observations, yet unmanned aerial vehicle (UAV) imagery is commonly acquired for a single task and retained by separate organizations. This study diagnoses the resulting reuse bottlenecks using an improved combinatorial weighting multi-criteria decision method (ICW-MCDM). Seven experts assessed four dimensions and 17 criteria. Data Rights (0.0912), Data Security (0.0823), Share Policy (0.0819), Data Description (0.0804), and Incentives (0.0706) received the highest integrated weights. A transparent raw-score comparator recovered the same five-item set, while bootstrap and leave-one-out checks supported a governance-oriented leading set with panel dependence for some criteria. After C1–C6 were excluded, Data Description, Reliable Data, Access Permissions, Service Facilities, Search and Discovery, and Data Citation and Provenance became the leading post-entry requirements. By 11 August 2026, a national directory platform had recorded 336,753 visits, fewer than ten formal applications, and two completed university research deliveries. The cases demonstrate small-scale matching, cross-institutional aggregation, and controlled delivery, but not general platform effectiveness or downstream GeoAI outcomes. The study separates governance entry from technical readiness and identifies governance and technical prerequisites for GeoAI-ready UAV data infrastructure. Full article

Automatic detection of grassland livestock from unmanned aerial vehicle (UAV) imagery can support livestock resource surveys and grazing management. However, compared with larger livestock such as cattle and horses, sheep in wide-field images are typically small, densely distributed, closely spaced, and easily confused with similar background textures, resulting in missed detections, background false positives, and duplicate detections. To address these challenges, this study presents DAFS-YOLO for UAV-based livestock detection in dense sheep scenes. A Dense-Aware Shallow Convolution module (DASConv) enhances shallow-feature discriminability for small-scale sheep in dense scenes, reducing missed detections and background false positives. A Local Spatial–Semantic Complementary Mapping module (LSCM) preserves richer shallow spatial information during feature propagation, improving small-object localization. Gaussian Soft-NMS optimizes the selection of highly overlapping candidate boxes in dense sheep regions and reduces duplicate detections. Experiments on a self-constructed UAV livestock dataset show that DAFS-YOLO achieves ${\mathrm{m}\mathrm{A}\mathrm{P}}_{50}$ and ${\mathrm{m}\mathrm{A}\mathrm{P}}_{50:95}$ values of 0.930 and 0.634, outperforming YOLOv11n by 4.4 and 4.8 percentage points, respectively. The corresponding sheep-class values are 0.922 and 0.564. With only 2.65 M parameters, the model also demonstrates good cross-dataset generalization on SheepCounter, providing an effective solution for intelligent UAV-based livestock monitoring. Full article

Driven by the escalating global agricultural workforce shortage and the urgent need to meet the “Zero Hunger” mandate, the automation of harvest–transport workflows has emerged as a cornerstone of Agriculture 4.0. This paper highlights the latest research progress in multi-robot collaborative logistics scheduling across highly unstructured farming environments, underpinned by cutting-edge spatial perception and digital twin frameworks. Initially, we summarize the technological leap from conventional 2D geometric mapping to multi-modal semantic 3D reconstruction—fusing light detection and ranging (LiDAR), unmanned aerial vehicle (UAV) imagery, and spatial data—to enable high-fidelity forward-looking predictions. The discussion then transitions to algorithmic advancements, emphasizing the shift from traditional centralized operations research to decentralized, data-driven approaches such as Multi-Agent Reinforcement Learning (MARL). We also explore micro-kinematic predictive control mechanisms and the growing integration of ecological sustainability metrics into routing models. To demonstrate practical engineering progress, multi-agent implementations are analyzed across three typical spatial settings: high-throughput continuous relays in open fields, global navigation satellite system (GNSS)-denied discrete routing in dense orchards, and close-proximity human–robot collaboration (HRC) in smart greenhouses. Finally, we identify the remaining barriers to the large-scale commercialization of Agricultural Cyber-Physical Systems (ACPS), such as the “Sim-to-Real” gap restricted by edge-computing capacities, unclosed economic loops, and HRC ethical dilemmas, offering a forward-looking roadmap for next-generation resilient agricultural networks. Full article

Cauliflower emergence rate and seedling growth are key indicators of field conditions and varietal potential. Traditional manual surveys are unsuitable for continuous monitoring across multiple varieties. This study integrates UAV RGB imagery with the DualSlim-YOLO model to estimate cauliflower emergence rates and monitor seedling growth. Built on YOLOv11, the model incorporates a lightweight feature extraction structure and an optimized detection-scale configuration. It reduces computational complexity while maintaining detection accuracy, thereby improving the efficiency of cauliflower seedling detection. DualSlim-YOLO achieved P, R, F1-score, mAP@0.5, and mAP@0.5:0.95 of 95.35%, 96.75%, 96.05%, 98.55%, and 86.65%, respectively. The number of parameters was reduced by 38.61%, while the inference speed increased by 22.16%, demonstrating good lightweight performance. Based on this model, UAV images of 171 cauliflower varieties acquired at 7, 21, and 28 d after transplanting were used for seedling detection and emergence rate estimation. In addition, 18 time-series seedling phenotypic traits were extracted, enabling a comprehensive quantitative evaluation of emergence dynamics and early-growth performance across multiple cauliflower varieties. This method effectively screens cauliflower varieties for high emergence rates, rapid emergence, and excellent seedling growth performance. It provides technical support for high-throughput, nondestructive seedling phenotyping and early germplasm screening under field conditions. Full article

Visible foliar damage in maize caused by fall armyworm(Spodoptera frugiperda)is commonly assessed through manual field scouting, a labor-intensive process that is difficult to scale across large planting areas. Multispectral and hyperspectral sensing can support more objective assessment but remain costly and impractical for routine field use. This exploratory case study evaluated whether low-cost red–green–blue (RGB) imagery can provide preliminary indicators of visible foliar damage associated with natural S. frugiperda infestation. RGB video was recorded in maize fields in Jerez, Zacatecas, Mexico, yielding seven field-acquired sequences and 302 extracted frames. The pipeline combined hue–saturation–value (HSV)-based foliar segmentation with four visible-spectrum indices—Excess Green (ExG), Excess Red (ExR), the Visible Atmospherically Resistant Index (VARI), and the Green Leaf Index (GLI)—an ExG-ratio damage threshold, and a 17-feature descriptor per frame used to train a Random Forest (RF) severity classifier. The study is positioned relative to RGB, Unmanned Aerial Vehicle (UAV)-based, deep learning, and multimodal approaches through its emphasis on traceability, low acquisition cost, and sequence-aware validation. Using the recovered canonical HSV/ExG-ratio pipeline, sequence-level mean damage ranged from 0.1086% to 0.4511%, with maximum frame-level damage up to 7.7401%. Severity labels were percentile-derived from the canonical damage index, yielding 100 Low, 99 Medium, and 103 High samples. Under a stratified frame-level split, the RF baseline reached 76.9% accuracy and a macro F1-score of 0.759. Under leave-one-sequence-out validation, performance decreased to 57.3% overall accuracy and 0.577 macro F1-score, indicating sequence-level dependence and supporting a conservative interpretation of classifier generalization. A zero-shot comparison using the Segment Anything Model (SAM) on a curated ten-frame-per-sequence subset produced higher damage estimates (SAM 1.30–11.67% versus HSV 0.82–2.61% on the same frames), suggesting HSV segmentation may under-detect pale or bleached tissue. These results provide preliminary, exploratory evidence that low-cost RGB indices can capture information associated with visible foliar damage in the studied recordings, without establishing agronomic validation, generalization beyond this dataset, or readiness for field deployment. Full article

Object detection in unmanned aerial vehicle (UAV) imagery is severely challenged by extremely small object scales, cluttered backgrounds, and pronounced foreground–background imbalance, which jointly degrade the accuracy of general-purpose detectors. This paper presents HSF-Net, a small-object detection network built upon YOLOv11s through three complementary enhancements. First, a Hierarchical Edge Enhancement Module (HEEM) employs orthogonal strip-convolution decomposition with zero-initialized residual fusion to enhance and re-weight the fine-scale edge and texture cues that are progressively attenuated in deep convolutional backbones. Second, a Sparse-Aware Feature Modulation (SAFM) module replaces concatenation-based fusion in the top-down neck pathway, coupling a sparse foreground gate with channel-wise scale modulation to confine cross-scale aggregation to object-bearing regions. Third, the detection head is restructured from {P3, P4, P5} to {P2, P3, P4}, introducing a high-resolution pathway for tiny targets while removing the original low-resolution P5 branch. On VisDrone, HSF-Net attains 46.4% mAP₅₀ and 28.5% mAP_50–95, exceeding the baseline by 7.1 and 4.8 percentage points, respectively, while reducing the parameter count from 9.4 million to 3.6 million. The model achieves an end-to-end throughput of 131.3 FPS on an NVIDIA RTX 3090, although the high-resolution P2 branch increases the computational cost to 39.5 GFLOPs. After dataset-specific training and evaluation on the markedly different TinyPerson benchmark, HSF-Net outperforms YOLOv11s by 5.2 percentage points in mAP₅₀, indicating that its relative performance advantage persists under a substantially different target-scale distribution and maritime background. Full article

Individual-tree-level information on crown distribution and morphology underpins forest inventory, biomass estimation, and carbon accounting. High-resolution Unmanned Aerial Vehicle (UAV) imagery resolves single-tree detail, but automated crown extraction remains difficult: fully supervised segmentation depends on costly pixel-level annotation that generalizes poorly, while the Segment Anything Model (SAM), though training-free, cannot locate trees on its own and existing SAM-based methods restore this ability only by adding task-specific training. We present an end-to-end, annotation-free toolkit for tree crown extraction and ecological analysis. A RetinaNet-based DeepForest detector produces coarse boxes; an adaptive module then removes duplicate boxes and non-vegetation false positives using an intersection-over-union rule and a global greenness index, converting noisy boxes into clean prompts; these prompts drive SAM to decode irregular crown masks without task-specific training; and geometric and texture features are extracted and grouped by principal component analysis and K-means clustering to map ecological patterns. We evaluated the toolkit on multi-biome imagery from the public OAM-TCD dataset. Because pixel-exact metrics are unstable at 10 cm resolution, where wind sway, shadow shift, and small labeling offsets are strongly amplified, we assessed accuracy under an absolute physical-distance tolerance. At a 2.0 m tolerance, consistent with the effective radius of a mature crown, the toolkit reached a precision of 91.25%, a recall of 86.40%, and an F1-score of 88.76%; bootstrap and Monte Carlo resampling confirmed these values are stable. Without manual annotation, it characterized more than 4700 individual crowns and recovered distinct vegetation patterns across geographic settings, offering a highly adaptable, low-cost baseline tool that demonstrates robust performance across the diverse multi-biome scenes within the OAM-TCD dataset. Full article

Accurate quantification of aboveground biomass (AGB) for sparse desert vegetation is fundamental for assessing regional carbon sink potential, yet large-scale data retrieval remains constrained by high spatial heterogeneity and severe background noise. Focusing on Haloxylon ammodendron communities in the Gurbantunggut Desert, this study integrated plot-level ground truth derived from Unmanned Aerial Vehicle Light Detection and Ranging (UAV-LiDAR) with multi-source satellite imagery (Sentinel-2 and Jilin-1) to evaluate AGB estimation accuracy and spatial distribution patterns across various feature combinations and employed four machine learning algorithms at a 10 m pixel scale. The Difference Vegetation Index (DVI) exhibited the strongest explanatory power for AGB spatial variance, whereas downsampled high-resolution textures induced feature redundancy. Among the evaluated algorithms, the Random Forest (RF) model driven solely by multispectral parameters achieved the optimal cross-scale mapping accuracy (R² = 0.72, RMSE = 1.32 t ha⁻¹). The total regional AGB storage was estimated to be approximately 6.99 × 10⁴ t, with low-density habitats (0.2–2.0 t ha⁻¹) occupying 90.39% of the area. This study confirms the feasibility of integrating UAV point clouds with multi-source satellite imagery for the large-scale retrieval of sparse shrub biomass, providing a quantitative basis for desert carbon management. Full article

Addressing the challenges posed by the small scale, diverse morphology, and severe occlusion of pest and disease targets on lotus leaves in complex aquatic environments—as well as the difficulty of existing methods in automatically quantifying disease severity—this paper proposes an approach for the identification of lotus leaf pests and diseases and quantitative grading of leaf spot disease severity. First, by integrating high-altitude canopy imagery captured by unmanned aerial vehicles (UAVs) with high-definition ground-level data, a multi-perspective dataset comprising object detection bounding box annotations and pixel-level segmentation annotations is constructed. Second, the YOLOv12n-DFFN object detection model is proposed; this model enhances interaction between deep and shallow features through a dynamic feature feedback mechanism, thereby improving the ability to localize disease targets against complex aquatic backgrounds. Finally, taking typical leaf spot disease as the subject, YOLOv12n-DFFN is used to detect and extract diseased leaf regions, while a VGG-UNet semantic segmentation model is employed to achieve pixel-level fine-grained segmentation of leaf and lesion areas. By calculating the ratio of lesion area to total leaf area, automatic quantitative grading of disease severity is realized. Experimental results show that YOLOv12n-DFFN achieved a precision of 93.58%, representing an improvement of 4.37 percentage points over the YOLOv12n baseline, with an mAP50 of 83.87%; the segmentation model attained an average intersection-over-union of 90.47%, and the overall accuracy of the two-stage framework for leaf spot disease severity grading reached 96.0%. Through a “detection first, segmentation second” two-stage strategy, this framework enables the intelligent identification and severity quantification of lotus leaf diseases in complex aquatic environments, providing an effective approach for the intelligent monitoring and precision management of aquatic crop diseases. Full article

Reliable inspection of insulator defects from unmanned aerial vehicle (UAV) remote sensing imagery is essential for the safe operation of power transmission systems. However, the task remains challenging due to fine-grained defect patterns, thin structures, large-scale variations, and complex backgrounds in aerial scenes. To address these issues, we propose Insulator-DETR, an end-to-end detection transformer specifically designed for UAV-based insulator defect inspection. The proposed framework preserves the set-prediction paradigm of DETR while introducing task-oriented modifications. In the encoder, a Parallel Attention MLP integrates multi-scale spatial perception and phase-aware token interaction to enhance the representation of fine textures and structural details. In the decoder, a Masked Linear Attention mechanism combines efficient global modeling with local contextual aggregation, enabling accurate localization of irregular defects. Furthermore, a convolutional feed-forward design is adopted to strengthen spatial interactions. Extensive experiments on two newly annotated UAV insulator-defect datasets and a public benchmarks demonstrate that Insulator-DETR consistently outperforms state-of-the-art detectors in both detection accuracy and recall. Full article

Unmanned aerial vehicle (UAV) imagery supports intelligent surveillance, environmental monitoring, traffic management, and infrastructure inspection. Yet aerial detection is difficult when objects are small, crowded, and observed at markedly different scales. Background clutter and illumination changes further weaken target cues and impair localization. We therefore propose HD-YOLO, a lightweight multiscale detector for small objects in UAV imagery. Its Multi-Dilation Shared Convolution Kernel (DSCK) extracts local texture and contextual information with shared dilated kernels. The Hybrid Dilated Bidirectional Feature Pyramid Network (HDFPN) reconstructs global and local cues before bidirectional aggregation, enabling high-resolution evidence to reach the prediction layers. The Efficient and Slim Head (ES-Head) combines shared operations with differential convolution to reduce cost and strengthen boundary-sensitive features. A joint ShapeIoU and Normalized Wasserstein Distance loss improves regression for small, irregular objects. Together, these components reduce missed detections in dense, cluttered scenes without relying on large model capacity. On VisDrone2019, HD-YOLO improves precision, recall, mAP50, and mAP50:95 over YOLOv8n by 6.9%, 7.2%, 8.2%, and 5.2%, respectively, while reducing parameters from 3.0 M to 0.9 M. Evaluations on TinyPerson and HIT-UAV also support its utility for tiny pedestrians and infrared aerial targets. HD-YOLO therefore improves small-object detection with a compact parameter footprint, while direct hardware benchmarks remain necessary to establish deployment efficiency. Full article

Cross-view geo-localization (CVGL) between unmanned aerial vehicle (UAV) imagery and satellite imagery is a key technique for autonomous UAV navigation in Global Navigation Satellite System (GNSS)-denied environments. However, most existing methods rely on energy-intensive Artificial Neural Networks (ANNs), making them difficult to deploy on resource-constrained edge computing platforms. Spiking Neural Networks (SNNs) provide a promising alternative for energy-efficient inference, but their application to CVGL still faces two challenges that remain insufficiently addressed. First, the isotropic computation used by existing SNN backbones is mismatched with the directional characteristics of spike activations. Spike activations tend to form oriented aggregation patterns along elongated geographic structures, and isotropic computation can therefore dilute directional signals. Second, the limited representational capacity of SNNs further increases the sensitivity during training optimization. However, the standard triplet loss adopts a static weighting strategy and assigns the same weight to all triplets that violate the margin constraint, which is unfavorable for learning from hard negatives. To address these challenges, we propose a framework with two core contributions. At the feature extraction level, the Directional Adaptive Convolution Module (DACM) processes spike feature maps by sequentially performing horizontal strip convolution and vertical strip convolution, thereby capturing a more complete geometric structure of directional spike clusters. At the training supervision level, we propose a Dual-dimensional Progressive Reweighting (DPR) loss, which jointly characterizes sample difficulty from pairwise difficulty and positive-pair quality difficulty. A learnable fusion parameter is used to adaptively balance these two types of difficulty information. Experimental results on the University-1652 and SUES-200 benchmarks show that the proposed framework, when equipped with the same representation learning head as its ANN counterparts, achieves competitive and, in many settings, superior performance. In terms of energy efficiency, its estimated theoretical energy consumption is over $8.8\times$ lower than that of published ANN methods under their original configurations. Under a more rigorous matched ANN control that shares the identical architecture, the estimated energy is reduced from 29.84 mJ to 6.36 mJ, an approximately $4.7\times$ reduction obtained at a cost of only 2.29 percentage points in R@1. Full article