Search Results (6,446)

Printed circuit board (PCB) defect inspection is critical for ensuring product reliability, yet it remains challenging due to the microscopic scale of defects and complex background patterns. To improve the localization of fine anomalies, this paper proposes EAS-DETR, an efficient and highly sensitive real-time end-to-end detector. First, we reconstruct the feature extraction backbone by introducing a novel C2f-EC module, which jointly models local textures and global structural dependencies. Second, an Adaptive Sparse Attention-based Intra-scale Feature Interaction (ASAFI) module is proposed to suppress background noise and focus the network’s attention on sparse defect regions. Finally, an optimized feature pyramid network, SGO-FPN, is designed to mitigate cross-scale feature misalignment and preserve high-resolution spatial details for small object localization. Experiments demonstrate that EAS-DETR achieves an mAP@0.5 of 93.0% and a 91.9% recall on a multi-source PCB dataset. The model outperforms mainstream YOLO variants and baseline RT-DETR models while maintaining a moderate parameter count of 14.6M and achieving a real-time inference speed of over 70 FPS. Furthermore, cross-domain validations on public benchmarks confirm its robust generalization capability for complex tiny object detection tasks. Full article

Electrified railways are major industrial electricity consumers. The Grid-Source-Storage-Vehicle (GSSV) system supports a more sustainable railway power supply by improving local renewable energy utilization, strengthening multi-source energy coordination, and promoting low-carbon development. However, existing rule-based energy management strategies (EMS) remain limited in their ability to support the efficient coordinated operation of the GSSV system. Moreover, under strong source-load fluctuations, conventional optimization-based EMS often fail to provide sufficiently reliable and responsive decision-making for real-time operation of GSSV systems. To address these issues, this paper proposes a real-time EMS based on a rule-guided enhanced non-dominated sorting genetic algorithm (RG-NSGA-II). First, based on the GSSV architecture, the operating modes of the system under different working conditions are systematically analyzed, and a corresponding rule-based EMS is designed. Then, a multi-objective optimization model considering system economic performance and grid power-intake fluctuation is formulated. Furthermore, a coordination mechanism between the rule-based EMS and the optimization EMS is developed. By embedding power commands generated by the rule-based EMS into the optimization EMS and regulating their activation through a time threshold, the proposed method improves the reliability, economic efficiency, and real-time performance of the EMS. Finally, the proposed method is validated, and the results show that the proposed real-time EMS ensures effective utilization of RE, improves power coordination efficiency and operational adaptability under fluctuating operating conditions, and delivers tangible environmental and economic sustainability benefits for electrified railway power supply systems. Full article

Traditional pedestrian inertial navigation (PDR) algorithms usually assume that the carrying mode of a smartphone is fixed and remains horizontal, while ignoring the significant impact of dynamic changes in the carrying mode on heading estimation, which is the core element of PDR algorithms. In practical application scenarios, pedestrians often change their way of carrying smart terminals (e.g., calling) according to their needs, corresponding to the difference in the heading estimation method; especially when the mode is switched, it will cause a sudden change in heading, which will lead to a significant increase in the localization error if it cannot be corrected in time. Existing smart terminal carrying mode recognition methods that rely on traditional machine learning or set thresholds have poor robustness; lack of universality, especially weak diagnostic ability for mutation; and can not effectively reduce the heading error. Based on these practical problems, this paper innovatively proposes a PDR framework that tries to overcome these limitations. Based on this research purpose, firstly, this paper classifies four types of common carrying modes based on practical applications and designs a CNN-LSTM hybrid model, which can classify the four common carrying modes in near real-time, with a recognition accuracy as high as 99.68%. Secondly, based on the mode recognition results, a multi-layer heading correction strategy is introduced: (1) introducing a quaternion-based universal filter (VQF) algorithm to realize the accurate estimation of initial heading; (2) designing an algorithm to accurately detect the mode switching point and developing an adaptive offset correction algorithm to realize the dynamic compensation of heading in the process of mode switching to reduce the impact of sudden changes; and (3) considering the motion characteristics of pedestrians walking in a straight line segment where lateral displacement tends to be close to zero. This study designs a heading optimization method with lateral displacement constraints to further inhibit the drifting of the heading caused by the slight swaying of the smart terminal. In this study, two validation experiments are carried out in two different environment—an indoor corridor and a tree shelter—and the results show that based on the proposed multi-layer heading optimization strategy, the average heading error of the system is lower than 1.5°, the cumulative positioning error is lower than 1% of the walking distance, and the root mean square error of the checkpoints is lower than 2 m, which significantly reduces the positioning error and shows the effectiveness of the framework in complex environments. Full article

Nonlinear model predictive control (NMPC) delivers high tracking accuracy for industrial processes but requires solving a nonlinear program at each sampling instant, limiting its applicability under tight real-time constraints. The Koopman operator provides a principled route to circumvent this limitation by embedding nonlinear dynamics into a higher-dimensional space where the evolution becomes linear, thereby reducing the online optimization to a convex quadratic program. This paper presents a Koopman-based MPC framework (K-MPC) that incorporates three algorithmic contributions. First, an adaptive radial basis function dictionary learning procedure selects lifting functions from process data, eliminating manual basis selection and improving approximation fidelity for systems with localized nonlinearities. Second, a recursive least-squares update rule adjusts the Koopman matrix online as new measurements arrive, enabling the controller to track slow parameter drifts without full model recomputation. Third, a tube-based constraint tightening strategy accounts for the residual linearization error, preserving recursive feasibility under bounded Koopman approximation mismatch. Simulations on a Van der Pol oscillator, a continuous stirred-tank reactor (CSTR), and a four-state Tennessee Eastman-inspired distillation column demonstrate that K-MPC achieves root-mean-square tracking errors within 11–16% of NMPC while reducing average per-step computation time by a factor of 14 to 18. The recursive update mechanism reduces prediction error by 80% compared to the fixed offline Koopman model when reactor feed concentration drifts by 15% from its nominal value. Ablation experiments confirm that adaptive dictionary learning and online updating each contribute measurably to closed-loop performance. Full article

The Qipanshan area in Northeastern China has diverse stand types and abundant forest resources, but extremely low resistance to external disturbances such as forest fires. Thus, improving the accuracy of understory fuel moisture content prediction is crucial for local forest fire prevention. This study focused on surface branch fuels in four typical stands (Larix gmelinii (Rupr.) Kuzen forest, Betula platyphylla Sukaczev forest, Pinus sylvestris var. Mongholica Litv. forest and cutover land) to evaluate the prediction and extrapolation performance of three hourly scale models (Nelson, Simard, and meteorological element regression models), and analyze their variations with slope positions and stand types, filling the gap in local hourly fuel moisture prediction model application. Results indicated that obvious spatial heterogeneity in fuel moisture content, closely affected by slope, fuel decay degree and microclimate, and thick, badly decayed branches had higher moisture content, with the highest in the Betula platyphylla forest and the lowest in cutover land. In terms of prediction accuracy, the Nelson model performed best, followed by the Simard model, while the meteorological element regression model was the poorest; predictions were more accurate in Pinus sylvestris var. mongholica forest and cutover land, and better on upper slopes than middle and lower slopes. For extrapolation capacity, the Simard model was optimal, followed by the Nelson model, while the meteorological element regression model was unfit for extrapolation due to excessive errors; extrapolation accuracy was best in cutover land and upper slopes. This study clarifies the applicability of the three models, providing methodological support for accurate real-time forest fire danger forecasting in the region. Full article

The high penetration of distributed renewable energy in rural microgrids imposes severe physical-layer fluctuations, weak information-layer communication, and limited computing-layer resources. These triple constraints create a fundamental tension: high-precision forecasting and real-time scheduling are required, yet edge devices face severe resource limitations. To resolve this, we present an edge-deployable energy management system (EMS) that achieves forecasting–scheduling co-optimization. We first propose an Adaptive Gated Dual-stream Network (AGDN), which employs a feature-dimension gated fusion mechanism to overcome the limitations of the local dependency strengths of Long Short-Term Memory (LSTM) and the global perception capabilities of Transformer models under volatile rural conditions. This approach achieves a Mean Absolute Percentage Error (MAPE) of 4.2% for load forecasting, outperforming baseline models by a significant margin. Next, we introduce a Prediction Uncertainty-Guided Quantum-Inspired Optimization (PUG-QIO) algorithm that adaptively maps prediction confidence intervals to quantum rotation angles, enabling deep integration of forecasting and scheduling and yielding an energy utilization rate of 93.2%. Finally, a Temporal Sensitivity-Aware Differentiated Pruning (TSADP) strategy is developed to maintain forecasting accuracy under a 63% parameter compression, overcoming the deployment barrier for high-precision models on edge devices. A 30-day field trial confirms that the proposed system meets the stringent rural requirements across four critical dimensions: forecasting accuracy, real-time responsiveness, lightweight architecture, and economic viability. Overall, the proposed system satisfies four key rural requirements: forecasting accuracy (MAPE = 4.2%), real-time response (≤10 s), lightweight deployment (memory < 500 MB), and economic viability (27.3% fuel cost reduction). Full article

As an innovative metaheuristic algorithm, Genghis Khan Shark Optimization (GKSO) faces challenges, including a tendency towards local optima and poor convergence speed and accuracy. To mitigate these limitations, an improved Genghis Khan shark optimizer (IGKSO) is proposed in this paper. A population partitioning method based on cosine similarity and fitness is introduced, where individuals are strategically assigned to different evolutionary phases: Disadvantaged populations are responsible for the foraging stage. By contrast, advantaged populations dominate the moving stage. In the moving stage, the base vector is randomly selected from multiple candidates, which ensures the evolutionary direction of the population while maintaining its diversity. An adaptive step-size mechanism is introduced to avoid boundary overflow problems. A subspace method is employed to prevent diversity loss during foraging. Additionally, in the hunting stage, a novel opposition-based learning strategy is proposed to moderate the tendency of converging to suboptimal solutions. Furthermore, during the self-protection phase, a criterion for assessing the diversity of the whole population is employed to monitor and supplement diversity in real time. The results of the CEC2017 and CEC2019 benchmark test sets reveal that IGKSO exhibits substantial advantages over the GKSO algorithm and eight other high-performance algorithms in terms of convergence speed and accuracy. Full article

This paper presents an edge-deployable vision-based framework for human–robot interaction using a xArm collaborative robot and a single RGB camera mounted on the robot wrist, and lightweight AI-based perception modules. The system enables intuitive, contact-free control by combining hand understanding and object detection within a unified perception–decision–control pipeline. Hand landmarks are extracted using MediaPipe Hands, from which continuous hand trajectories, static gestures, and dynamic gestures are derived. Task objects are detected using a YOLO-based model, and both hand and object observations are mapped into the robot workspace using ArUco-based planar calibration. To ensure stable robot motion, the hand control signal is smoothed using low-pass and Kalman filtering, while dynamic gestures such as waving are recognized using a lightweight LSTM classifier. The complete pipeline runs locally on edge hardware, specifically NVIDIA Jetson Orin Nano and Raspberry Pi 5 with a Hailo AI accelerator. Experimental evaluation includes trajectory stability, gesture recognition reliability, and runtime performance on both platforms. Results show that filtering significantly reduces hand-tracking jitter, gesture recognition provides stable command states for control, and both edge devices support real-time operation, with Jetson achieving consistently lower runtime than Raspberry Pi. The proposed system demonstrates the feasibility of low-cost edge AI solutions for responsive and practical human–robot interaction in collaborative industrial environments. Full article

The homeodomain-leucine zipper (HD-Zip) transcription factor family is conserved in land plants and is critical for regulating growth, development, and stress responses. Flax (Linum usitatissimum L.) is an economically valuable dual-purpose crop valued for its high nutrition and notable drought tolerance; however, its HD-Zip gene family has not been systematically characterized. In this study, a comprehensive genome-wide analysis was performed to identify and characterize the HD-Zip family in flax. A total of 34 LuHD-Zip genes were identified, which were unevenly distributed across 15 chromosomes and exhibited substantial variation in physicochemical properties. The encoded proteins ranged from 200 to 372 amino acids in length, with molecular weights of 22.7–40.3 kDa and theoretical isoelectric points (pI) of 4.49–9.46. All LuHD-Zip proteins were predicted to be hydrophilic and localized to the nucleus. Phylogenetic analysis divided these proteins into two major subfamilies (Group 1 and Group 2), a classification strongly supported by conserved gene structures and motif compositions, implying potential functional redundancy within each group. Gene duplication analysis revealed that segmental duplication events (29 pairs) were the primary drivers of family expansion. Comparative syntenic analysis further indicated that the LuHD-Zip gene family has remained relatively conserved throughout evolution. Promoter cis-element analysis identified multiple regulatory elements associated with hormone signaling and abiotic stress responses, suggesting complex transcriptional control in response to environmental stimuli. Expression profiling via quantitative real-time PCR (qRT-PCR) demonstrated that LuHD-Zip genes exhibit tissue-specific expression patterns and are differentially regulated by various phytohormone treatments and abiotic stresses. This study provides the first genome-wide characterization of the HD-Zip gene family in flax, offering valuable insights into its evolution and potential functions. These findings establish a solid foundation for future functional investigations of the LuHD-Zip gene family. Full article

The relentless miniaturization of semiconductor components and Printed Circuit Boards (PCBs) has rendered Automated Optical Inspection (AOI) of micro-defects a critical bottleneck in modern manufacturing and metrology. While in-line inspection systems offer economically viable and scalable quality control solutions, they impose stringent constraints on both inference latency and detection robustness—particularly for diminutive, sparsely distributed defects (e.g., mouse bites, pinholes) amidst complex, repetitive circuit topologies. To bridge this gap, we present a semantic-enhanced FCOS framework specifically engineered for micro-defect inspection. Our approach introduces two synergistic innovations: (1) a Semantic-Guided Upsampling Unit (SGU) that adaptively reweights channel–spatial features to reconcile the semantic disparity between shallow textural details and deep contextual representations; and (2) a Sparse Center-ness Calibration (SCC) module that enforces high-confidence, spatially sparse supervision to sharpen localization precision and suppress false positives. The SGU is integrated within a Progressive Semantic-Enhanced Feature Pyramid Network (PSE-FPN) that extends multi-scale representations to stride-4 (P2) resolution, while the SCC module is embedded directly into the detection head. Comprehensive evaluations on MS COCO and the real-world DeepPCB dataset validate the efficacy of our design. On COCO, our model achieves 41.8% AP with real-time throughput of 28 FPS on a single NVIDIA 1080Ti GPU. A lightweight variant further attains 41.6% AP at 42 FPS, accommodating high-throughput production environments. For PCB defect detection, the framework delivers 98.7% mAP@0.5, substantially outperforming contemporary detectors. These results demonstrate that semantics-aware, lightweight architectures enable scalable, real-time quality assurance in semiconductor manufacturing. Full article

Machine learning (ML) is transforming the analysis of biomolecular data, holding significant promise for improving the efficiency and accuracy of microscopy image analysis and for studying the dynamics of molecules in live cells. As data-driven approaches continue to evolve, they may eventually replace traditional statistical methods that rely on conventional analytical methods. This review examines and critically analyses the state of the art of ML techniques as applied to various levels of data supervision in the analysis of dynamic single-molecule datasets obtained using superresolution optical microscopy. Collectively encompassed under the umbrella of “nanoscopy”, these methods currently comprise targeted techniques such as stimulated emission depletion (STED) microscopy and stochastic techniques like single-molecule localization microscopies (SMLMs), comprising photoactivated localization microscopy (PALM), DNA points accumulation for imaging in nanoscale topography (DNA-PAINT) microscopy, and minimal fluorescence photon flux (MINFLUX) microscopy. These techniques all enable the imaging of subcellular components and molecules beyond the diffraction limit, and some are additionally capable of studying their dynamics in real time, as reviewed here, using several ML techniques that facilitate motion analysis in two or three dimensions with qualitative and quantitative characterisation in the live cell. It is expected that the growing use of learning-based approaches in biological microscopy data processing will dramatically increase throughput and accelerate progress in this rapidly developing field. Full article

Automated detection of suspicious human activities in complex and crowded environments remains a critical challenge in modern surveillance systems due to high false-positive rates, poor contrast and generalization across diverse scenes. We propose a GM_CNN3D Model for the classification of suspicious activity based on a Deep Fused Feature Block (DFFB) framework that integrates handcrafted spatial descriptors (PCA-HOG and Motion-HOG) with deep spatiotemporal features extracted from 3D Convolution Neural Network (3D-CNN). Motion regions are first localized using a Gaussian Mixture Model (GMM), after which handcrafted and deep features are concatenated in a dimensionality-normalized fusion stage, followed by a fully connected layer and softmax classification. The system is evaluated on five diverse and publicly available datasets: Violent Crowd, Hockey Fight, Kaggle Fight, Movies Fight, and Custom Annotated YouTube Clips, achieving up to 99.12% accuracy, 98.7% F1-score, and a ROC-AUC of 0.992, outperforming state-of-the-art CNN, LSTM, and SlowFast models. All datasets include real world scenarios with varying lighting, crowd density, and camera viewpoints, with annotations created manually where unavailable. The proposed method demonstrates robust cross-scene performance, enabling automated alarming and reduced false positives in real-time security operations. Full article

Despite over two decades of advancement in object detection, achieving high accuracy for small target detection in practical applications remains an unresolved challenge. This paper proposes a novel small-object detection model to address this issue. The model incorporates three key innovations: first, the RCSOSA module, which optimizes feature information transmission through dynamic channel interaction and multi-scale feature coordination; second, the HFPN module, a three-branch multi-scale feature fusion network that integrates local and global features by combining CNN and Transformer architectures to enhance semantic details; and third, the NWD-CIoU loss function, which dynamically adjusts the weights of NWD and CIoU losses based on the training phase. Experimental results on the COCO dataset demonstrate that our model improves detection accuracy by 4% over YOLOv11 and achieves state-of-the-art performance among mainstream models while maintaining a real-time inference speed of no less than 60 FPS. Furthermore, validation on the VisDrone dataset confirms the model’s strong generalization capability. The proposed algorithm significantly enhances small target detection accuracy, effectively mitigating a critical limitation in current practical object detection applications. Full article

Bacterial biofilms are structured microbial communities enclosed within a self-produced extracellular polymeric substance matrix composed of polysaccharides, proteins, extracellular DNA, and lipids. This matrix promotes adhesion, structural stability, and the development of heterogeneous microenvironments that restrict antimicrobial penetration and shield bacteria from host immune responses. As a result, biofilms are major contributors to chronic, recurrent, device-related, and difficult-to-treat infections, posing a major challenge for clinical management and antimicrobial stewardship. This review summarizes current understandings of biofilm biology, its clinical relevance, including the stages of biofilm development, the composition and protective roles of the matrix, and the physiological heterogeneity that arises during maturation. It also examines key mechanisms underlying biofilm tolerance and resistance, such as limited antibiotic diffusion, and sequestration, enzymatic inactivation, efflux pump upregulation, persister cell formation, and horizontal gene transfer. In addition, it highlights important clinical settings in which biofilms are implicated, including cystic fibrosis, chronic wounds, osteomyelitis, implant- or device-associated infections, and breast implant illness, in which persistent implant-associated biofilms and the resulting chronic inflammatory milieu have been hypothesized to contribute to local and systemic manifestations in a subset of patients. The review further discusses conventional and emerging approaches for biofilm detection alongwith real-time monitoring. Biofilm-associated infections remain difficult to eradicate because persistence is driven by multiple interconnected protective mechanisms. Effective management therefore requires integrated strategies that combine accurate detection with multifaceted therapies, including antibiotics alongside matrix-disrupting enzymes, quorum-sensing inhibitors, bacteriophages, metabolic reactivators, and nanotechnology-based delivery systems. Advances in multi-omics and system-level modeling will be essential for developing next-generation strategies to prevent, monitor, and treat biofilm-associated disease. Full article

The integration of high-penetration distributed photovoltaics (PV) into distribution networks triggers frequent voltage limit violations, fluctuations, and increased network losses. To address the limited communication infrastructure inherent in medium-voltage distribution networks, this paper employs PV inverters as fast-response voltage regulation devices and proposes a real-time distributed voltage control strategy specifically for such networks. Firstly, a distribution network communication topology and voltage regulation architecture based on adjacent asynchronous communication are established. A reactive power-voltage tracking regulation method at PV grid connection points is introduced, utilizing the division and equivalence of voltage regulation feeder segments. By partitioning the distribution network into feeder segments centered around individual PV units, rapid reactive power-voltage tracking regulation based on local and neighboring information is achieved. Secondly, a three-stage cascaded real-time distributed voltage control strategy integrating both reactive power regulation and active power curtailment is designed. Within each regulation stage of this strategy, a voltage estimation process is embedded, enabling dynamic evaluation of the regulation effectiveness and adaptive determination for transitioning between stages. Finally, the proposed strategy is applied to modified IEEE 33-node and IEEE 69-node test systems. Simulation results verify the effectiveness and superiority of the proposed method in improving voltage quality and reducing network losses. Full article