Search Results (11,355)

This study addresses the challenges of sparse long-range point clouds, complex background interference, and inconsistent localization quality in 3D object detection for intelligent driving in open-pit mines. A camera–LiDAR multimodal detection method based on Voxel R-CNN is proposed. We introduce a Multimodal Focal Sparse Voxel Enhancement module that combines Focal Sparse Convolution with shallow visual features to guide voxel importance prediction and selective sparse propagation. An Intersection over Union (IoU)-Aware Quality and Geometry Refinement Head is further designed to improve the localization accuracy and ranking reliability of 3D proposals. Experimental results show that the proposed method achieves BEV mAP@0.40 and 3D mAP@0.40 values of 63.43% and 61.23%, respectively, outperforming the strongest comparison method, MambaFusion, by 5.37 and 7.61 percentage points. In the 60–80 m range, the average translation error is reduced to 0.41 m, while the inference speed reaches 27 FPS. These results demonstrate that the proposed method improves the detection and localization of distant sparse objects in complex open-pit mine environments while maintaining real-time inference capability. Full article

Humans acquire meaningful language by storing perceptual categories, category-word mappings, and conditional IF–THEN rules in rich, cross-domain, multimodal contexts. Crucially, structured cross-domain input that pairs visual context and language (text) appears to be fundamental to this process, enabling individuals to acquire, for example, the lexicon. The present study investigates whether Large Language Models (LLMs) can learn words when trained on structured cross-domain input rather than text-only exposure. This paper incorporates a controlled Fictitious-Animal Paradigm featuring 32 creatures and 32 pseudowords, divided into two phases: single-animal and dual-animal scenes. Single-animal contexts entailed storing categories and mappings, while dual-animal environments involved storing logical constraints and performing inference. The authors propose a computational model centred on four capacities for context-based word learning: storing perceptual categories, category-word mappings, and the Mutual Exclusivity rule (formalised as a conditional IF-THEN statement), and retrieving stored information to assign novel labels through inference. The evaluation demonstrates that structured cross-domain input enables LLMs to exhibit behaviour consistent with conditional IF–THEN rules, allowing them to infer and acquire novel words based on current visual contexts. These findings suggest that, with this input that integrates context and text, LLMs display adaptive, real-time human-like inferential reasoning in word learning. Full article

Reliable crop mapping is essential for understanding agricultural practices, optimizing resource use, and analyzing rural dynamics, while also supporting modelling and sustainable agriculture planning. However, obtaining 10 m crop distribution maps remains challenging in Mediterranean regions, where data availability is often limited and landscapes are fragmented. The main European land use dataset, Corine Land Cover (CLC), lacks both the crop specificity required for accurate crop differentiation and the temporal frequency needed for timely monitoring. This study addresses these limitations by implementing the IOTA² automated chain in Sardinia (Italy), to create a large-scale crop map specifically targeting Mediterranean crops. The methodology leverages open-source satellite imagery with supervised machine learning, using the 2018 Land Parcel Identification System (LPIS), CLC, and Urban Atlas dataset for training. We compared two nomenclatures, detailed (32 classes) versus simplified (25 classes), testing each across three training sample sizes (10%, 50%, and 100%). Results indicate that the simplified nomenclature (N25) provided more robust performances, achieving an overall accuracy (OA) of 0.77 with full sampling, compared to 0.61 for the detailed version. These OA values refer to the subset of reference polygons held out from the reference data for independent pixel-level validation. Moreover, a final map was produced using the entire reference dataset for training and evaluated through zonal area agreement. Mapping showed high performance for specific crops like rice, citrus, and grapevine, while classes such as cereals and fruit trees presented classification challenges due to high fragmentation of the landscape and irregular crop-distribution patterns. Despite these challenges, this work delivers a 10 m spatial resolution reproducible framework that enhances thematic details of current European datasets. By running as a single automated processing chain rather than a sequence of manually executed steps, it offers a scalable solution for rapid, annual crop monitoring in complex, data-scarce Mediterranean environments. Full article

Multivariate time series forecasting requires modeling both hierarchical temporal dynamics and complex inter-variable dependencies, a dual requirement that often degrades predictive performance and incurs high computational costs in standard Transformer architectures. Unlike current channel-independent models that ignore vital cross-variable synergies, or dense-attention frameworks that suffer from quadratic computational noise, our approach extracts structurally sparse dependencies. To address these specific limitations, this study introduces the FCA-Transformer. The proposed framework integrates a Feature Pyramid Network (FPN) to isolate macroscopic trends from high-frequency localized fluctuations via hierarchical downsampling. Concurrently, a structured Transformer-based Cross-Attention (TCA) mechanism employs Dimensional Segmentation with Weighting (DSW) and a Two-Stage Attention (TSA) layer to map topological variable interactions, effectively extracting robust cross-variable pathways and mitigating distributional noise. Extensive empirical evaluations across three real-world multivariate benchmarks (ETTh1, Electricity, and Exchange Rate) demonstrate that the FCA-Transformer achieves an average reduction of up to 4.39% in MSE and 5.11% in MAE compared to leading baselines. These findings indicate that the proposed architecture successfully reconciles multi-scale feature extraction with lightweight dependency modeling, enhancing structural generalization and providing a scalable framework for real-time temporal analysis in complex industrial environments. Full article

As a meta-heuristic inspired by human student learning behaviors, the original Student Psychology-Based Optimization (SPBO) suffers from insufficient exploitation of historical population records, simplistic individual interaction patterns and high risk of falling into local optima. This work develops an enhanced SPBO (ESPBO) embedding three dedicated learning mechanisms. The adaptive knowledge-accumulation learning component imports personal historical best, global elite and population-mean information into position update formulas to sustain coherent search trajectories and enhance convergence precision. The multi-level peer collaborative learning module categorizes agents into excellent, intermediate and under-performing groups based on fitness values. Customized learning rules are configured for each group to enable diverse information sharing: elite individuals expand promising search regions, medium-level agents learn from counterparts, and inferior individuals move toward high-quality candidates. The progressive examination feedback component dynamically modulates search intensity by measuring the fitness improvement of each individual, so as to better balance global exploration and local exploitation. Comparative numerical experiments are carried out on CEC2017 and CEC2022 benchmark test suites against multiple advanced meta-heuristic algorithms. Results indicate that ESPBO exhibits outstanding accuracy and robustness on unimodal, multimodal, hybrid and composite test functions. To explore its real-world applicability, ESPBO is adopted for mobile-robot path-planning simulations under multi-scale grid maps. Simulation results from 20 × 20, 40 × 40 and 60 × 60 environments illustrate that ESPBO stably produces collision-free trajectories, outperforming comparative algorithms in path length, smoothness and safety performance. It is demonstrated that the three embedded learning mechanisms substantially strengthen the optimization capacity of vanilla SPBO, and ESPBO possesses considerable application potential for numerical optimization as well as mobile robot path-planning scenarios. Full article

The aquatic products industry has experienced rapid development due to the growing global demand for high-quality proteins. However, conventional production models remain constrained by multifaceted impediments, including disease outbreaks, environmental stressors, microbial contamination, and operational inefficiencies in processing, quality control, and logistics. Artificial intelligence (AI) offers targeted methodological solutions to these challenges. From a functional perspective, this review categorizes artificial intelligence into four major types: perception, prediction, control, and generation, and systematically evaluates its application progress in aquaculture, processing, quality inspection, and traceability fields. Perception AI constitutes the data acquisition and digitization layer, utilizing computer vision, sonar, and multimodal fusion technologies to establish digital mappings from environmental parameters to biological indicators. Prediction AI employs machine learning and deep learning algorithms to transform historical datasets into quantitative forecasts regarding water quality dynamics, disease risks, and production trends. Control AI translates decision-making protocols into precise, autonomous regulatory actions for aquaculture environments and processing workflows through fuzzy logic and model-based predictive control. Generation AI leverages large language models and generative adversarial networks to demonstrate innovative capabilities in data augmentation, solution optimization, and virtual simulation. Collectively, these applications optimize core production processes while significantly enhancing product quality, processing efficiency, safety management, and traceability systems. Future research directions will prioritize the development of robust, interdisciplinary AI technologies with superior integration capabilities. Full article

Inshore synthetic aperture radar (SAR) ship detection remains challenging because strong coastal clutter, speckle noise, and degraded target responses frequently result in false alarms and missed detections of small vessels. Moreover, local scattering-related responses and high-level semantic information exhibit different characteristics across network stages, making it difficult for a unified feature-learning framework to fully exploit their complementarity. To address these issues, we propose a Scattering–Semantic Collaborative Network based on an asymmetric dual-branch DINO architecture, termed S-DINO. Specifically, a scattering-guided local feature reconstruction module establishes correlations between dispersed high-response regions and selected dominant response centers to improve the representation of fragmented target structures and enhance small-ship representation. Furthermore, a Semantic–Scattering Dual-Driven Query Injection strategy combines normalized feature-space response magnitude with semantic confidence to guide candidate reference-box initialization and reduce the selection bias caused by strong coastal activations. Experiments on the SSDD and HRSID inshore subsets demonstrate that, compared with the baseline DINO, S-DINO improves mAP@50 by 5.6 and 11.5 percentage points and F1-score by 11.4 and 8.6 percentage points, respectively. These results indicate the effectiveness of collaboratively exploiting scattering-related response cues and semantic information for ship detection in complex inshore environments. Full article

Continuous objective assessment of pig motor condition, including recovery after anesthesia, requires the joint analysis of posture, locomotion, transitions between states, and short-term adverse events. This study aimed to develop and algorithmically evaluate a multi-camera pipeline for the quantitative analysis of simulated recovery states in a fully synthetic virtual environment. Procedural generation was used to produce 150,000 images annotated with fifteen anatomical keypoints while varying the pose, size, and appearance of the model, camera viewpoints, illumination, environment, and image post-processing parameters. The YOLO11m-pose model was used for two-dimensional pose estimation, after which observations from three synchronized cameras were combined using weighted triangulation. The reconstructed three-dimensional trajectories were processed using a quality-control system, a finite-state machine, and temporal rules for detecting falls, prolonged immobility, and convulsion-like movements. After repartitioning the dataset by three-dimensional pose index, the retrained YOLO11m-pose model achieved a keypoint ${\mathrm{mAP}}_{50--95}$ of 0.8968, a precision of 0.9985, and a recall of 0.9986 on the independent pose-level test set. During the processing of a 15-minute three-camera sequence, 98.993% of 405,000 reconstructions satisfied the geometric acceptance criteria, and the median reprojection error was 2.420 pixels. On a first manually annotated 5-minute synthetic sequence, the state machine achieved a strict frame-level accuracy of 91.20%, a macro-F1 score of 90.10%, and Cohen’s $\kappa$ of 0.860; accuracy outside $\pm 1$ s neighborhoods of state transitions was 97.43%. On a second 5-minute synthetic evaluation sequence with exact Unity-world geometric ground truth, direct three-dimensional evaluation yielded an MPJPE of 31.86 mm at 99.993% landmark coverage. A limited temporal assessment on this second sequence detected both of the two lying-derived prolonged-immobility reference intervals; fall and convulsion-like-movement events were not independently annotated. All training, validation, and end-to-end evaluation data were generated in a virtual environment; therefore, these results establish algorithmic feasibility within the synthetic domain but do not establish performance on real animals. Full article

Hybrid cyber–physical systems (CPS) integrate cloud services used in enterprise environments with operational technology (OT), which controls physical processes. However, most threat models applied to such systems implicitly assume that an adversary necessarily causes any loss of process availability. This perspective is reflected in both the STRIDE model and the MITRE ATT&CK for ICS knowledge base, which primarily focus on adversarial activities. As a result, they do not explicitly account for a scenario that is becoming increasingly relevant in hybrid architectures: the intentional shutdown of a physical process by the organization defending the system. The loss of availability resulting from the activation of protective mechanisms is not, in itself, a new problem. The concept of a spurious trip has been recognized in safety engineering for decades and is addressed in standards such as IEC 61511. Related dependencies are also considered within the STPA-Sec methodology. Therefore, the objective of this work is not to introduce a new type of threat, but rather to demonstrate that this phenomenon is not adequately represented in widely used threat-modeling taxonomies that are primarily attacker-centric. In addition, a specific trust boundary within the hybrid architecture at which this problem manifests itself is identified. This makes it possible to incorporate the phenomenon into the risk analysis of systems in which a compromise of the IT layer alone can ultimately lead to the shutdown of physical processes. The proposed threat model is based on trust-boundary analysis. The reference hybrid architecture was divided into seven trust boundaries, and each STRIDE category was subsequently mapped to the corresponding MITRE ATT&CK techniques for ICS, based on the trust boundary crossed by a given attack scenario. The resulting threat vectors were then ranked using the fundamental metrics defined in CVSS v4.0. The attack vector was derived from the trust boundary crossed by each scenario and, where applicable, was correlated with published CVE vulnerability assessments. The model was validated against four widely documented industrial cybersecurity incidents: Stuxnet, Triton, Industroyer, and Colonial Pipeline. Full article

The offspring of domesticated Pinctada maxima exhibited various physiological and microbial adjustments to the complex and dynamic conditions of coastal environments. To support the restoration of P. maxima genetic resources and explore the molecular mechanisms underlying these phenotypic responses, we conducted a comparative analysis of the intestinal transcriptome and microbiota of wild parental and domesticated generations. Each sample generated an average of 43,600,525 clean reads, which were mapped to the P. maxima reference genome with mapping rates ranging from 63.91% to 79.48%, and a total of 3007 differentially expressed genes (DEGs) were subsequently identified. Gene Ontology analysis revealed that the DEGs were enriched in organic acid metabolism, and Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis revealed that the DEGs were enriched in glycosphingolipid biosynthesis and xenobiotic metabolism via cytochrome P450. Microbiota profiling revealed significant compositional shifts at phylum and genus levels, with increased alpha diversity in F1; dominant phyla transitioned towards Spirochaetota and Bacteroidota, and functional predictions pointed to enhanced metabolic and immune capacities. Quantitative validated the up-regulation of immune genes (PmHR96h, PmIAP) and down-regulation of calcium-signaling genes (PmCaM, PmHSP90), consistent with RNA-seq data. Collectively, these coordinated transcriptomic and microbial alterations reflect a multifaceted host–microbiome adaptive response to nearshore conditions. Our findings provide valuable molecular markers and microbial indicators for selective breeding and health monitoring, offering a scientific basis for improving the resilience of P. maxima aquaculture under changing environmental conditions. Full article

By combining digital sensing, network connectivity, data-driven analyses, cloud services, and automated controls, smart farming has been increasingly adopted in agricultural production. Although these technologies have improved the precision and efficiency of farm management, they also increase cybersecurity exposure as agricultural facilities are connected to external networks, platforms, and remote-control environments. This review seeks to clarify why cybersecurity should be considered a fundamental requirement in smart farming and details the major system components, cybersecurity risks, and network design considerations required for secure operation. This review first explains the concept and application scope of smart farming, and then examines how sensors, communication networks, gateways, control systems, data platforms, user interfaces, cloud infrastructure, and physical support systems contribute to farm management and cybersecurity exposure. The review also emphasizes that smart farming differs from ordinary information systems because digital data and control commands can directly affect physical processes, such as irrigation, ventilation, heating, nutrient supply, and livestock management. Based on these cyber-physical characteristics, the review summarizes the key architectural considerations for reducing cybersecurity risks, including network segmentation, data and command flow mapping, gateway and wireless security, remote access management, cloud access control, device inventory, logging, monitoring, resilience, and local fail-safe operation. Overall, ensuring cybersecurity in smart farming requires an integrated approach that protects not only data and accounts but also the reliability and continuity of agricultural production. Full article

The Italian coastal zone is a dynamic and fragile environment that is characterized by high ecological vulnerability and subjected to strong anthropogenic pressures that significantly alter its natural morphology and environmental balance. In this scenario, port infrastructures represent about 35% of the Italian artificial coast and have a significant influence on coastal dynamics and seabed integrity. This study proposes a new methodology based on polygon-based mapping to calculate the spatial extent and distribution of the transformation areas associated with port development. This allows us to define a new specific indicator, named the Port Pressure Indicator, to assess the impact of port infrastructures. The proposed analysis highlights that port-related pressures on the marine-coastal zone are driven by two main factors: large-scale ports, which account for more than half of the total transformation areas, and a widespread network of minor ports, which are responsible for the systematic fragmentation of coastal continuity. By providing a standardized measurement of transformation areas, this approach supports coastal spatial planning and monitoring. Moreover, it provides a baseline for assessing spatial pressure and supporting integrated coastal zone management assessments. Full article

Vision-based hand gesture recognition (HGR) systems are predominantly developed for frontal camera viewpoints, whereas smart-home cameras are often ceiling-mounted, creating a viewpoint-induced domain gap. To investigate this issue, we collected and manually annotated CeilGest, an 18-class ceiling-view hand gesture dataset comprising 156,282 annotated frames from 68 participants recorded in distinct domestic environments. We then systematically evaluated frontal-to-ceiling transfer using YOLO-based detectors trained on HaGRID and compared their performance with an in-domain CeilGest-trained model. On identical ceiling-view footage, the frontal-trained YOLOv8n produced approximately 24× more class-to-class misclassified frames than the in-domain model (486 vs. 20 across 27,000 frames); this large paired difference remained evident when temporal dependence within gesture holds was taken into account. The effect was strongly class-dependent, with AP decreasing by up to 5.5 percentage points for the worst-affected gesture, while the aggregate same-architecture mAP50 difference was 0.6 percentage points. Across five YOLOv8 variants evaluated on frontal HaGRID, mAP50 remained at 0.995, supporting selection of the lightweight YOLOv8n for edge deployment. The complete ceiling-view HGR pipeline was implemented on Raspberry Pi 5 using NCNN and Jetson Orin Nano using TensorRT. Mean inference latency was 74.69 ± 4.72 ms and 12.72 ± 0.12 ms, respectively. During a 10 min continuous Raspberry Pi 5 test, mean inference latency increased by 17.6% and junction temperature reached 90.8 °C with thermal throttling. Power consumption and INT8 inference were not evaluated. Overall, the results show that training–deployment viewpoint consistency is a major consideration for ceiling-mounted HGR and establish an in-domain supervised baseline relative to frontal-only training without domain adaptation. Full article

Facade debonding may lead to falling components and pose safety risks in dense urban environments, but infrared thermal responses may also arise from non-defect facade components and environmental conditions. Conventional facade inspection methods are often labor-intensive, hazardous, and difficult to integrate into digital maintenance workflows. To support safer and more efficient facade inspection and maintenance information management, this study develops an engineering-oriented inspection and management framework that integrates unmanned aerial vehicle infrared thermography, intelligent defect recognition, visual result verification, and defect information management. A UAV-based infrared data acquisition scheme was established, and a self-constructed dataset containing 1035 thermal images was developed for the detection of suspected facade debonding regions and common thermal interference sources, including windows, air-conditioning units, and signage. A lightweight detection model was embedded as the recognition engine of the framework to balance detection reliability and deployment efficiency under practical inspection conditions. Experimental results show that the proposed method achieved an mAP@0.5 of 87.8%, with 1.64 million parameters and 4.2 GFLOPs, indicating its potential for rapid preliminary facade screening under the tested computing configuration. Beyond model evaluation, an application platform was developed to support infrared image and video input, automatic detection, result visualization, statistical analysis, and defect record storage. The proposed framework demonstrates the potential of combining UAV infrared inspection and digital management tools for preliminary facade screening and inspection documentation, providing supporting information for subsequent engineering review and maintenance planning. Full article

Background and Objectives: Non-invasive mechanical ventilation is the first-line treatment for managing acute respiratory failure. However, patient variability and complex pulmonary mechanics complicate therapy adjustments, frequently leading to ventilator-induced lung injuries. This study aims to propose and define a simulation platform and decision support prototype, named VENTILA2, to optimize pressure-controlled ventilation strategies. Methods: The system integrates a bicompartmental series model of the respiratory system incorporating severity-stratified physiological profiles of chronic obstructive pulmonary disease and acute respiratory distress syndrome, and it is coupled with a Mamdani fuzzy inference system. This architecture maps inspiratory time adjustments based on pressure errors and their derivatives across predefined clinical profiles within a scenario-based feedforward parameter-mapping framework. Results: Evaluated through quantitative operational verification across all profiles and proof-of-concept case studies, the platform successfully recreates complex clinical scenarios, accurately simulating phenomena such as accelerated lung emptying in severe acute respiratory distress syndrome and air trapping in moderate chronic obstructive pulmonary disease. Conclusions: VENTILA2 provides a controlled simulation environment for evaluating pathology-specific ventilatory configurations across simulated profiles prior to clinical implementation, though it remains an early-stage prototype whose clinical effectiveness, safety, and robustness remain to be rigorously evaluated. Full article