Search Results (3,342)

Maize kernel damage detection is critical for grain quality assessment and post-harvest processing. However, existing deep learning methods struggle to balance accuracy, model complexity, and multi-class recognition under real-world conditions. To address these issues, this paper proposes YOLOv11n_DEG, an improved lightweight detection model based on YOLOv11n. The model uses the first ten pretrained layers as a feature extractor, replaces standard convolutions in the backbone with depthwise separable convolutions to reduce parameters, integrates an Efficient Channel Attention (ECA) module to enhance feature representation, and employs a dual-dropout strategy in the classification head to mitigate overfitting and improve generalization. Additionally, to account for potential discrepancies in damage characteristics between the obverse and reverse sides of maize kernels, a dual-sided synchronous image feature fusion method is introduced. The output layer classifies five target categories for multi-class damage detection. On an independent test set, the proposed model achieves a precision of 92.5%, a recall of 89.6%, and a mean average precision (mAP) of 93.4%, outperforming the original YOLOv11n by 6.2%, 4.0%, and 3.7%, respectively, while reducing parameter count and computational complexity by 32.9% and 6.5%. To validate practical deployability, a custom testbed with dual-camera synchronous acquisition and geometry-based matching was developed. On this platform, the model with dual-sided fusion achieves a single-side recognition accuracy of 94.1% and a dual-sided recognition accuracy of 89.3%, with an average detection time of 0.8 s per batch. These results demonstrate that YOLOv11n_DEG provides an accurate and practical solution for intelligent maize kernel damage detection, with strong potential for real-world deployment in grain inspection systems. Full article

Systematic measurement biases are a persistent source of degradation in airborne radar target tracking. Online bias estimation is especially difficult for single-platform operations because only one measurement stream is available, and the target state and radar biases are coupled in the same nonlinear observation model. Under weakly observable geometries, directly augmenting bias states into a recursive filter may lead to slow convergence, while fixed or overly frequent batch optimization may inject updates from weakly informative windows. To address this problem, this paper proposes a Bias Subspace Information-Guided MAP-EKF (maximum a posteriori–extended Kalman filter) (BI-MAP-EKF) for online bias estimation in single-platform airborne radar. A nine-dimensional augmented state jointly describes target motion and range, azimuth, and elevation biases. A posterior Cramér–Rao lower bound (PCRLB) is constructed for this state, where the EKF prior, windowed radar measurements, and process noise propagation are jointly considered. The bias subspace is then extracted by Schur complement, and the resulting azimuth bias PCRLB is used to decide whether the MAP update is reliable enough for EKF injection. The scheduler also includes a cooldown interval and a maximum-window safeguard, which respectively limit excessive updates under informative maneuvers and prevent indefinite waiting under weak geometries. Fifty-run Monte Carlo simulations under different maneuvering conditions show that the proposed scheduler is particularly effective in weakly informative geometries, where it improves azimuth bias and horizontal position estimation while reducing the number of accepted MAP refinements compared with the tested fixed-period MAP and moving-horizon estimation (MHE) baselines. In stronger maneuvering scenarios, it provides a competitive accuracy–cost trade-off rather than uniformly outperforming aggressive MHE in every channel. Full article

During the fabrication of prestressed concrete beams, the quality and positional accuracy of the laid corrugated ducts (or prestressing ducts) directly influence the load-bearing capacity and durability of the beams. However, traditional manual inspection is inefficient, highly subjective, and difficult to achieve full coverage. To address this problem, this paper proposes an automated detection method that integrates improved YOLOv11-based pose estimation, robust curve fitting, and image stitching techniques. The method automatically identifies duct positions and evaluates laying quality. By incorporating the SE channel attention mechanism and the SPPFCSPC multi-scale pooling module, the SC-YOLOv11 model is developed, which significantly enhances the detection accuracy of slender corrugated pipe key points in environments with dense rebar occlusion. The RANSAC algorithm is employed to fit curves to the predicted key points, effectively suppressing the influence of outliers. Furthermore, the SIFT algorithm is used for precise stitching of drone-captured segmented images, which are then transformed into a unified front orthographic coordinate system of the entire box girder via perspective transformation, enabling accurate reconstruction of the corrected 2D layout of corrugated ducts across the full beam. Ablation experiments using 5-fold cross-validation demonstrate that SC-YOLOv11 improves mAP₅₀ and mAP_50–95 by 2.6% and 1.2%, respectively, with statistical significance (paired t-test, p < 0.01). The model achieves a per-image inference time of 6.37 ms, with 4.34 M parameters and 8.1 GFLOPs, meeting real-time requirements. In a 30 m prefabricated box girder field application, the measured section trajectory fitting curves of the corrugated ducts were compared with the design alignment, successfully identifying two abnormal locations where the laying deviation exceeded the allowable threshold. Cross-validation with on-site inspector records shows that over 92% of the measurement points agree within ±10 mm. This method achieves a fully automated analysis chain from key point detection and curve fitting to deviation quantification, providing an efficient, non-contact, and traceable intelligent tool for quality control of bridge prestressed systems. Full article

Robotic teleoperation is a core enabling technology spanning remote surgery, industrial inspection, and virtual-reality applications. Despite growing deployment, operator experience assessment still relies almost exclusively on post hoc subjective questionnaires, which preclude real-time monitoring and adaptive intervention. Here, the wearable IMU chain is considered not only as a command interface but also as an implicit sensing channel for operator state. We test whether end-effector kinematics generated by the IMU-to-robot mapping contain information about ten post-task workload and user-experience dimensions, comprising NASA-TLX-inspired workload scales together with usability, responsiveness, realism, intuitiveness, and perceived performance. A secondary analysis of a publicly available dataset (16 participants, 144 motion recordings, simulated UR10e arm) is conducted through a three-stage pipeline: bivariate correlation analysis (Pearson and Spearman), multivariate regression (10 model families, 16 feature-set combinations, Leave-One-Subject-Out validation), and binary classification (median-split). Statistical validity is assessed via 1000-permutation nested testing. Target-specific regression models reach $R^2\ge 0.50$ on seven out of 10 subjective dimensions, with a peak of $R^2=0.787$ for usability; permutation testing confirms significance for eight out of 10 targets. Binary classification achieves AUC $\ge 0.75$ on nine out of 10 targets, with three dimensions reaching perfect AUC. SHAP analysis identifies temporal irregularity and distributional shape descriptors as the dominant kinematic explanatory families. These results support the feasibility of kinematics-based inference of operator experience and provide an offline proof of concept toward future real-time adaptive teleoperation systems. Full article

With the continuous penetration of renewable energy and power electronic equipment into modern power systems, the occurrence frequency of composite power quality disturbances has increased significantly. The accurate classification of various composite disturbances under strong noise remains a critical technical challenge. The existing single time–frequency transformation methods cannot simultaneously capture transient time-domain details and fine frequency-domain features of steady-state harmonics, while mainstream deep learning classification networks contain redundant parameters and introduce excessive computational overhead, failing to meet the real-time deployment requirements of power edge terminals. To address these limitations, a lightweight Coordinate Attention ResNet network named ResNet–LCA is proposed based on the dual-modal time–frequency fusion of the Continuous Wavelet Transform and Short-Time Fourier Transform. First, the two transforms are implemented separately to generate two groups of complementary time–frequency maps, which are concatenated along the channel dimension to fully extract the coupling features between the steady-state harmonics and the transient impulses. Second, a Haar wavelet subband mean aggregation module is designed for dimensionality reduction with negligible information loss. This module eliminates the channel redundancy introduced by the multimodal fusion and reduces the overall computational overhead at the input stage. Finally, a lightweight residual network integrated with Coordinate Attention is constructed, with Grouped Half-Convolution adopted to compress the model parameters. CA offsets the feature attenuation induced by the lightweight structural design and further improves the model’s noise immunity. A simulation verification was carried out on a simulated dataset covering 25 types of single and superimposed composite disturbances. At a signal-to-noise ratio of 20 dB, the proposed method achieved an average classification accuracy of 97.92%, with only 5.32 M total parameters and a single-sample GPU inference latency of 0.33 ms. Compared with standard ResNet-18 under 20 dB noisy conditions, the total parameter volume was reduced by 52.7%, the inference latency was shortened by 0.13 ms, and the classification accuracy was improved by 0.60 percentage points. The proposed method achieves coordinated optimization of classification accuracy, noise immunity and inference efficiency, and it can provide lightweight technical support for online intelligent power quality monitoring at the edge nodes of microgrids and islanded power systems. Full article

This study presents a multidisciplinary investigation of the fluvial evolution of the northern meander of the Calore River at Cellarulo locality, near Benevento (southern Italy). The research integrates archaeological evidence, geological and geomorphological data, historical cartography, remote sensing imagery and geophysical surveys in a Geographic Information System (GIS) environment. Multi-temporal analysis of historical maps, aerial photographs and satellite images from 1824 to 2022 allowed the reconstruction of channel migration patterns and the identification of abandoned meanders and paleochannel traces. Stratigraphic data derived from boreholes revealed the presence of a channel of the Calore River dated at least in the Bronze Age (3900 years ago), abandoned in the nineteenth century. Geoelectrical investigations provided detailed information on subsurface resistivity anomalies, highlighting the presence of buried structures and possible ancient anthropogenic features located at shallow depths between 1 and 1.5 m. The combined interpretation of geomorphological, archaeological and geophysical data demonstrates significant data on the unveiling of an ancient river channel and its abandonment during the last 150 years, suggesting a strong interaction between natural fluvial dynamics and human occupation. The results confirm the effectiveness of an integrated multidisciplinary approach for reconstructing fluvial landscape evolution and for identifying buried archaeological and geomorphological features in complex floodplain environments. Full article

Consumer sentiment shocks rarely remain confined to their economy of origin. Adopting a systems-thinking perspective, this study treats the equity markets of thirteen advanced economies as one interconnected socio-technical system, bounded by its environment. It maps how unexpected shifts in consumer confidence propagate across it between 2015 and 2025. Rather than isolating a single channel, the analysis examines the system as a whole, where a social subsystem of household sentiment interacts with a technical subsystem of market infrastructure. Sentiment shocks are identified as the unexpected component of the OECD Composite Consumer Confidence Index, and the dependency structure linking markets is estimated through return-based networks. The analysis combines the Diebold-Yılmaz connectedness framework, Granger-causal contagion testing, network centrality measures, and panel estimation with cross-sectionally consistent standard errors. Total connectedness reaches 81.6 percent, confirming a densely integrated system in which the Euro-area core acts as the principal return transmitter; sentiment-shock contagion, once corrected for multiple testing, is sparse rather than pervasive. A small set of economies occupies structurally central positions, yet the small-sample centrality diagnostic provides no robust evidence that threshold-network centrality predicts VAR-based net spillover roles. The findings refine the standard assumption that central nodes are necessarily the main propagators of systemic disturbance and offer concrete guidance for cross-border financial monitoring. This guidance is structural rather than a real-time monitoring signal since it derives from a full sample rather than a rolling or live analysis. Full article

Highly reflective surfaces often cause intensity saturation in captured fringe images, leading to phase errors and inaccurate 3D reconstruction. High-dynamic-range (HDR) fringe projection profilometry is an effective solution to this problem, but existing methods usually require additional image acquisition or auxiliary calibration, which limits their applicability to high-speed online inspection. In this work, we propose a phase-optimization-guided progressive adaptive fringe projection method for HDR 3D measurement. First, quality-guided RGB phase fusion is used to fuse reliable phase information from RGB channels, reducing unreliable phase regions and identifying residual overexposed areas. Second, local phase repair restores pixels that cannot be directly mapped in overexposed regions, thereby establishing stable camera–projector correspondence. Third, an adaptive optimal projection intensity is estimated through local intensity fitting to adjust the brightness of highly reflective regions. These steps are embedded into the frequency-varying phase-unwrapping process of conventional digital fringe projection, progressively suppressing saturation-induced phase distortion without additional projected/captured images or calibration operations. Experiments on a highly reflective blade and a white plaster bust demonstrate that the method suppresses saturation without extra image acquisition. Comparative results show high measurement accuracy and low missing ratios with fewer images. Full article

Reliable perception of young apple fruits in natural orchards is a prerequisite for automated thinning and intelligent orchard management, yet remains difficult in real field conditions due to small fruit size, dense distribution, branch–leaf occlusion, background similarity, and severe illumination degradation at night. This study presents FruitDet, a lightweight multi-module detector designed for robust day–night young apple fruit detection in complex orchard environments. A field dataset was established in a high-density apple orchard in Aksu, Xinjiang, covering daylight and low-light night-time scenes with diverse occlusion, scale, and illumination variations. To improve detection robustness without sacrificing computational efficiency, FruitDet combines three complementary mechanisms: an inverted-bottleneck-based multi-scale feature enhancement module for preserving small-fruit details, a channel–spatial attention module for suppressing foliage and illumination interference, and a lightweight Transformer-based context module for modeling long-range dependencies between fruits and surrounding orchard structures. In daytime scenes, FruitDet achieved 91.904% precision, 77.557% recall, 83.254% mAP50, and 66.427% mAP50–95; in night-time scenes, it maintained 90.107% precision, 75.135% recall, 80.544% mAP50, and 64.719% mAP50–95. Compared with mainstream detectors including YOLOv5n, YOLOv8n, YOLO11n, YOLO26n, Faster R-CNN, RT-DETR, and RT-DETRv2, FruitDet consistently delivered higher accuracy across lighting conditions. Ablation, visualization, public-dataset testing, and edge-deployment experiments verified that the proposed modules jointly improve small-object representation, background discrimination, low-light robustness, and real-time applicability. With 2.960 M parameters, 3.726 G FLOPs, and approximately 180 FPS, FruitDet offers a practical and efficient visual perception approach for Young fruit monitoring was conducted under both daytime and night-time orchard conditions covered in this study. All-weather orchard monitoring and robotic young-fruit thinning. The shareable data are available Full article

Pilot contamination attacks (PCAs) pose a great physical-layer threat to modern multiple-input–multiple-output (MIMO) systems by exploiting channel reciprocity to corrupt channel state information estimation. While existing literature acknowledges this vulnerability, precise parametric threat models that capture the transition from active channel estimation poisoning to data exploitation remain scarce. This article addresses this gap by developing a novel, physical-layer parameterized PCA framework structured as a two-stage operational attack: channel estimation poisoning and adaptive information contamination. We formulate an algorithmic attack strategy that systematically manipulates base transceiver station precoding and beamforming weights. This formulation allows us to quantify the precise degradation of the system through the lens of the confidentiality, integrity, and availability (CIA) triad, specifically mapping the adversary’s security gains against legitimate users’ signal-to-noise ratio degradation. Finally, we leverage this parametric threat model to outline a qualitative roadmap of actionable detection vectors and structural mitigation strategies. The proposed algorithmic attack serves as an evaluation benchmark and an analytical baseline for conceptualizing resilient architectures in emerging physical-layer security frameworks. Full article

This paper presents a numerical investigation of dispersion-managed dense wavelength division multiplexing (DWDM) transmission systems incorporating a non-midway optical phase conjugator (OPC) under randomly distributed residual dispersion per span (RDPS). Unlike conventional studies assuming ideal symmetric configurations, this work considers more realistic scenarios with asymmetric OPC placement and random dispersion distribution. To ensure the reliability of the analysis, simulations were performed for 100 different random RDPS patterns. A 960 Gb/s DWDM system consisting of 24 channels operating at 40 Gb/s was modeled using the nonlinear Schrödinger equation solved by the split-step Fourier method. To analyze the impact of OPC location, two asymmetric configurations (23-27 and 27-23), were compared. System performance was evaluated using eye-opening penalty (EOP) and timing jitter (TJ). The results show that OPC location has a significant impact on compensation efficiency, with the 27-23 configuration providing overall better performance than the 23-27 configuration. Although randomly distributed RDPS does not always outperform uniform or deterministic dispersion maps, certain random patterns achieve comparable or even superior compensation performance. Through extensive statistical evaluation across five independent random seeds, including Pearson correlation and regression analysis, we observed consistent structural tendencies in the RDPS distribution that enhance compensation efficacy. Specifically, in the 23-27 structure, a high correlation with a ‘half-cycle sin’ profile was preliminarily observed to be beneficial, whereas the 27-23 structure showed sensitivity to both ‘half-cycle sin’ and ‘one-cycle sin’ profiles. These findings suggest that maintaining antipodal-symmetry, even in stochastic environments, provides a stable probabilistic advantage for signal compensation. While we advise a cautious interpretation regarding the universal applicability of these results, the study offers valuable design insights and is expected to facilitate greater flexibility in the design of future high-capacity optical networks. Full article

This study investigates the high-temperature pyrolysis pathways and product distribution of the battery immersion coolant HFE-7300 (C₇H₃F₁₃O) within a simulated thermal runaway environment. Using a tube furnace system combined with GC-MS analysis across a temperature range of 300–800 °C (residence time of 3 s), the thermal stability and cracking evolution were evaluated. Experimentally, HFE-7300 exhibits low initial decomposition at 400 °C with a pyrolysis rate of 5.84%, which rapidly scales up to 48.72% at 500 °C, and reaches a near-complete degradation of 98.46% at 800 °C. Qualitative product characterization identified C₂H₄, C₂F₄, C₃F₆ C₄F₈, and C₅H₃F₉O as the primary species evolved. To map the micro-scale degradation trajectories, a reaction network comprising 12 elementary pathways was constructed via density functional theory (DFT) calculations at the B3LYP/6-311+G(d,p) level. Using the TST method, we calculated the reaction rate constants for the main decomposition pathways. Analysis reveals that the C4–C5 bond scission pathway (R6) serves as the predominant initial decomposition channel, yielding C₅H₃F₉O and CF₂=CF₂ as the definitive primary products. These findings provide baseline thermodynamic data and critical safety insights for the engineering design of immersion-cooled battery thermal management systems. Full article

Energy system optimization seeks to utilize multiple energy sources efficiently under technical, economic, and environmental constraints. The increasing integration of renewable energy and the need for sustainable operation have made the optimal planning and management of hybrid energy systems crucial. Classical optimization methods, including Linear Programming, Nonlinear Programming, and simulation-based models, often face limitations when addressing high-dimensional and nonlinear problems. This study introduces a deep learning–based surrogate modeling framework for sizing the components of hybrid renewable energy systems. Initially, Particle Swarm Optimization (PSO) is employed to determine the optimal component sizes for a large number of synthetically generated hourly solar irradiance and load profiles. These optimal solutions are then used as target labels. The associated annual time-series data are transformed into multi-channel Data Map (DMAP) images, which serve as inputs for convolutional neural networks (CNNs). After training, the CNN models are capable of directly estimating the required number of photovoltaic (PV) panels, inverter capacity, and battery units from the DMAP images, eliminating the need to perform the iterative PSO optimization during the prediction stage. Various convolutional neural network architectures, including ResNet, DenseNet121, RegNet, ConvNeXt, EfficientNet, SqueezeNet, MobileNet, and InceptionV3, were evaluated for this multi-output regression task. The results indicate that ResNet and DenseNet121 achieve the best performance, while ConvNeXt provides strong results with a modern architectural design. Among the evaluated models, DenseNet121 achieved coefficients of determination (R²) of 0.934, 0.988, and 0.947 for predicting the sizes of the PV array, inverter, and battery bank, respectively. These results correspond to an average prediction accuracy of approximately 90.6%. ResNet produced similar performance, with its highest R² value reaching 0.983 for inverter sizing. Lightweight networks such as SqueezeNet and MobileNet demonstrate notable effectiveness for resource-constrained systems, whereas InceptionV3 underperforms in leveraging its multi-scale architecture. These results demonstrate that, once the models have been trained, deep learning–based surrogate models can generate sizing decisions comparable to those obtained using PSO with only a fraction of the computational effort. As a result, they provide a fast and practical alternative to conventional iterative optimization methods for component sizing in smart grid and sustainable energy planning applications. Full article

Aiming at the problems of low facial expression recognition accuracy, high computational complexity, and difficulty in edge deployment caused by variable illumination, diverse head poses, severe partial occlusions, and the long-tail distribution of expression categories in complex power business hall scenarios, this paper presents an engineering-oriented lightweight dynamic perception facial expression recognition method for complex power business hall scenarios. The proposed method adopts an end-to-end joint face detection and expression classification framework built upon the anchor-free CenterNet architecture. In terms of technical implementation, this paper primarily focuses on the integration and adaptation of existing techniques to meet scenario-specific requirements; integrates a specially designed Multi-Scale Fusion Deformable Large Kernel Attention (MSF-DLKA) module to enhance multi-scale perception of subtle expression deformations under varying poses; designs a Task-Aware Dynamic Detection Head (TADDH) with decoupled spatial-channel attention to separately adapt the feature requirements of localization and classification subtasks; directly employs the off-the-shelf Label-Distribution-Aware Margin loss (LDAM Loss) to alleviate class imbalance; and combines three established compression techniques—structured pruning, quantization-aware training, and knowledge distillation—into a progressive lightweight pipeline for edge deployment. Experiments on two public micro-expression datasets, CASME II and SAMM, as well as a self-built electric power business hall scenario dataset (PBHD), show that the proposed method achieves mAP@0.5 of 93.2%, 93.9%, and 91.7%, and macro-F1 of 0.929, 0.936, and 0.914 on the three datasets, respectively, while using only 13.5 M parameters and attaining a single-image inference time of 28.9 ms on the NVIDIA Jetson AGX Orin. The paper provides a feasible solution for real-time expression perception in resource-constrained scenarios such as power business halls. Full article

To address the challenges in personnel detection in underground coal mines, such as uneven lighting, dust occlusion, cluttered backgrounds, and significant scale variation of targets, this paper proposes a lightweight object detection method named YOLO-SCC. Based on the lightweight detection network YOLO26n, the SCC lightweight enhancement structure, composed of SPDConv, CBAM, and ContextAggregation, is constructed to improve feature representation and target recognition capability in complex underground environments while striving to maintain model compactness. Specifically, SPDConv optimizes the downsampling process to reduce the loss of feature details, which helps preserve edge and texture information of miners under low-light conditions, at long distances, or at small scales. CBAM adaptively weights features from both channel and spatial dimensions, enhancing the model’s focus on the main body of miners and suppressing irrelevant interference from light reflections, equipment structures, and dust noise. The ContextAggregation module aggregates richer contextual semantic information, strengthening the model’s discriminative ability for personnel targets under occlusion, dense distribution, and complex backgrounds. Experimental results show that with only a modest increase in parameters (from 2.38 MB to 2.88 MB) and computational cost (from 5.2 GFLOPs to 6.0 GFLOPs), YOLO-SCC achieves precision of 88.8%, recall of 82.2%, mAP50 of 85.3%, and mAP50-95 of 51.2%. These represent improvements of 6.1, 0.9, 0.1, and 3.3 percentage points, respectively, over the baseline YOLO26n. The results suggest that YOLO-SCC primarily improves false-positive suppression and localization quality in complex underground environments while maintaining a lightweight model scale. Full article