Search Results (341)

To address the bottlenecks of traditional jaw crusher liner wear detection—high safety risks, insufficient precision, and limited full-range analysis—this paper proposes a non-contact, high-precision wear analysis method based on binocular vision and deep learning. At its core is the integration of the state-of-the-art FoundationStereo zero-shot stereo matching algorithm, following scenario-specific adaptations, into the 3D reconstruction of industrial liners for wear analysis. A novel wear quantification methodology and corresponding indicator system are also proposed. After calibrating the ZED2 binocular camera and fine-tuning the algorithm, FoundationStereo achieves an Endpoint Error (EPE) of 0.09, significantly outperforming traditional algorithms. To meet on-site efficiency requirements, a “single-view rapid acquisition + CUDA engineering acceleration” strategy is implemented, reducing point cloud generation latency from 165 ms to 120 ms by rewriting kernel functions and optimizing memory access patterns. Geometric accuracy verification shows a Mean Absolute Error (MAE) ≤ 0.128 mm, fully meeting industrial measurement standards. A complete process of “3D reconstruction–model registration–quantitative analysis” is constructed, utilizing three core indicators (maximum wear depth, average wear depth, and wear area ratio) to characterize liner wear. Statistical results—such as an average maximum wear depth of 55.05 mm—are highly consistent with manual inspection data, providing a safe, efficient, and precise digital solution for the predictive maintenance and intelligent operation and maintenance (O&M) of liners. Full article

This work presents an experimental evaluation of a high-speed analog-to-digital conversion method based on passive reference waveform crossings combined with time-to-digital converter (TDC) time-tagging. Unlike conventional level-crossing event-driven analog-to-digital converters (ADCs) that require dynamically updated digital-to-analog converters (DACs), the proposed architecture compares the input waveform against a broadband periodic sampling function without active threshold control. Crossing instants are detected by a high-speed comparator and converted into rising and falling edge timestamps using a multi-channel TDC. A commercial ScioSense GPX2-based time-tagger with 30 ps single-shot precision was used for validation. A range of test signals—including 5 MHz sine, sawtooth, damped sine, and frequency-modulated chirp waveforms—were acquired using triangular, sinusoidal, and sawtooth sampling functions. Stroboscopic sampling was demonstrated using reference frequencies lower than the signal of interest, enabling effective undersampling of periodic radio frequency (RF) waveforms. The method achieved effective bandwidths approaching 100 MHz, with amplitude reconstruction errors of 0.05–0.30 RMS for sinusoidal signals and 0.15–0.40 RMS for sawtooth signals. Timing jitter showed strong dependence on the relative slope between the acquired waveform and sampling function: steep regions produced jitter near 5 ns, while shallow regions exhibited jitter up to 20 ns. The study has several limitations, including the bandwidth and dead-time constraints of the commercial TDC, the finite slew rate and noise of the comparator front-end, and the limited frequency range of the generated sampling functions. These factors influence the achievable timing precision and reconstruction accuracy, especially in low-gradient signal regions. Overall, the passive waveform-crossing method demonstrates strong potential for wideband, sparse, and rapidly varying signals, with natural scalability to multi-channel systems. Potential application domains include RF acquisition, ultra-wideband (UWB) radar, integrated sensing and communication (ISAC) systems, high-speed instrumentation, and wideband timed antenna arrays. Full article

Object detection in adverse weather remains challenging due to the simultaneous degradation of visibility, structural boundaries, and semantic consistency. Existing restoration-driven or multi-branch detection approaches often fail to recover task-relevant features or introduce substantial computational overhead. To address this problem, DLC-SSD, a lightweight degradation-aware framework for detecting robust objects in adverse weather environments, is proposed. The framework integrates image enhancement and feature refinement into a single detection pipeline and adopts a hierarchical strategy in which global and local degradations are corrected at the image level, structural cues are reinforced in shallow high-resolution features, and semantic representations are refined in deep layers to suppress weather-induced noise. These components are jointly optimized end-to-end with the single-shot multibox detection (SSD) backbone. In rain, fog, and low-light conditions, DLC-SSD demonstrated more stable performance than conventional detectors and maintained a quasi-real-time inference speed, confirming its practicality in intelligent monitoring and autonomous driving environments. Full article

The rapid advancement of high-resolution remote sensing technology has significantly expanded observational capabilities in the oil and gas sector, enabling more precise identification of petroleum infrastructure. Remote sensing now plays a critical role in providing real-time, continuous monitoring. Manual interpretation remains the predominant approach, yet is plagued by multiple limitations. To overcome the limitations of manual interpretation in large-scale monitoring of upstream petroleum assets, this study develops an end-to-end, deep learning-driven framework for intelligent extraction of key oilfield targets from high-resolution remote sensing imagery. Specific aims are as follows: (1) To leverage temporal diversity in imagery to construct a representative training dataset. (2) To automate multi-class detection of well sites, production discharge pools, and storage facilities with high precision. This study proposes an intelligent monitoring framework based on deep learning for the automatic extraction of petroleum-related features from high-resolution remote sensing imagery. Leveraging the temporal richness of multi-temporal satellite data, a geolocation-based sampling strategy was adopted to construct a dedicated petroleum remote sensing dataset. The dataset comprises over 8000 images and more than 30,000 annotated targets across three key classes: well pads, production ponds, and storage facilities. Four state-of-the-art object detection models were evaluated—two-stage frameworks (Faster R-CNN, Mask R-CNN) and single-stage algorithms (YOLOv3, YOLOv4)—with the integration of transfer learning to improve accuracy, generalization, and robustness. Experimental results demonstrate that two-stage detectors significantly outperform their single-stage counterparts in terms of mean Average Precision (mAP). Specifically, the Mask R-CNN model, enhanced through transfer learning, achieved an mAP of 89.2% across all classes, exceeding the best-performing single-stage model (YOLOv4) by 11 percentage points. This performance gap highlights the trade-off between speed and accuracy inherent in single-shot detection models, which prioritize real-time inference at the expense of precision. Additionally, comparative analysis among similar architectures confirmed that newer versions (e.g., YOLOv4 over YOLOv3) and the incorporation of transfer learning consistently yield accuracy improvements of 2–4%, underscoring its effectiveness in remote sensing applications. Three oilfield areas were selected for practical application. The results indicate that the constructed model can automatically extract multiple target categories simultaneously, with average detection accuracies of 84% for well sites and 77% for production ponds. For multi-class targets over 100 square kilometers, manual detection previously required one day but now takes only one hour. Full article

Background/Objectives: Pulsed field ablation (PFA) is increasingly used for pulmonary vein isolation (PVI). One of the emerging single-shot PFA catheters is the variable-loop circular catheter (VARIPULSE™, Biosense Webster, Inc.) which is fully integrated into a three-dimensional mapping system. However, the evidence for the feasibility of ablation of non-pulmonary vein targets is still limited using the VARIPULSE catheter. In this study, we summarize our experience in PVI and mapping/ablation of non-pulmonary vein sites in patients with atrial fibrillation (AF) and complex atrial substrate and arrhythmias using the VARIPULSE catheter. Methods: All patients with paroxysmal or persistent AF who underwent catheter ablation using the VARIPULSE catheter were retrospectively included. PVI was performed in all patients. Spontaneous or inducible atrial flutters were mapped and ablated. Empiric lines were performed at the operator’s discretion. Acute outcomes and complications were analyzed. Results: the study included 60 patients; 25 (41.6%) were females and mean age was 67.15 ± 9.01 years. Thirty four (60%) had persistent AF and six (10%) patients had atrial flutter as the initial rhythm during the index procedure. All patients had PVI using the PFA as per protocol. Most of the patients (76.7%) had non-pulmonary vein ablation sites; posterior wall isolation was performed in 25 (41.7%) patients, roof line in 9 (15%) patients, anterior line in 16 (26.7%) patients, cavotricupsid isthmus in 11 (18.3%) patients and superior vena cava isolation in two (3.3%) patients. Overall, 27 patients had atrial flutters during the index procedure that were mapped and ablated using the VARIPULSE catheter. All had termination of atrial flutter except for one patient. Major complications were not detected. Conclusions: Mapping and ablation of atypical atrial flutter and non-pulmonary vein targets are feasible and safe using the VARIPULSE platform. Full article

Satellite segmentation is vital for spacecraft perception, supporting tasks like structural analysis, fault detection, and in-orbit servicing. However, the generalization of existing methods is severely limited by the scarcity of target satellite data and substantial morphological differences between target satellites and training samples, leading to suboptimal performance in real-world scenarios. In this work, we propose a novel diffusion-based framework for few-shot satellite segmentation, named DiffSatSeg, which leverages the powerful compositional generalization capability of diffusion models to address the challenges inherent in satellite segmentation tasks. Specifically, we propose a parameter-efficient fine-tuning strategy that fully exploits the strong prior knowledge of diffusion models while effectively accommodating the unique structural characteristics of satellites as rare targets. We further propose a segmentation mechanism based on distributional similarity, designed to overcome the limited generalization capability of conventional segmentation models when encountering novel satellite targets with substantial inter-class variations. Finally, we design a consistency learning strategy to suppress redundant texture details in diffusion features, thereby mitigating their interference in segmentation. Extensive experiments demonstrate that our method achieves state-of-the-art performance, yielding a remarkable 33.6% improvement over existing approaches even when only a single target satellite image is available. Notably, our framework also enables reference-based segmentation, which holds great potential for practical deployment and real-world applications. Full article

Conventional wavefront sensors face challenges when detecting frequency-domain information. In this study, we developed a high-precision, and fast multi-spectral wavefront detection technique based on neural networks. Using an etalon and a diffractive optical element for spectral encoding, the measured pulses were spatially dispersed onto the sub-apertures of the Shack-Hartmann wavefront sensor (SHWFS). We employed a neural network model as the decoder to synchronously calculate the multi-spectral wavefront aberrations. Numerical simulation results demonstrate that the average calculation time is 21.38 ms, with a root mean squared (RMS) wavefront residual error of approximately 0.010 μm for 4-wavelength, 21st-order Zernike coefficients. By comparison, the conventional modal-based algorithm achieves an average calculation time of 102.98 ms and wavefront residuals of 0.090 μm. Remarkably, for 10-wavelength analysis, traditional centroid algorithms fail; this approach maintains high simulation accuracy with the RMS wavefront residual error below 0.016 μm. The proposed approach significantly enhances the measurement capabilities of SHWFS in multi-spectral and single-shot wavefront detection, particularly for single-shot spatio-temporal characterization in ultra-intense and ultra-short laser systems. Full article

Tomato is one of the most important vegetable crops worldwide, with about one quarter of the yearly production of fresh fruits dispatched to the processing industry. Paste, canned tomatoes, and sauces represent the three leading categories. Background/Objectives: The metabolic profile of processed tomatoes can be modified by several production steps, affecting the nutritional and sensory profile of the finished product. Despite this, a detailed metabolomic profiling of transformed tomatoes is currently missing. The goal of this investigation is to provide qualitative metabolomic profiling of tomato purees with two main advances: first, the use of a more sustainable analytical approach based on a single extraction protocol and one-shot analysis for multiple information retrieval on different compound classes; second, the achievement of a curated database consolidated over a wide collection of commercial samples representative of the Italian market. Methods: A non-selective ethanol extraction was applied to collect the main polar metabolites followed by untargeted high-resolution MS/MS analysis and software-based compound identification. Results: A list of more than five hundred features was collected and assigned to specific compounds or compound groups with different confidence levels. The results confirmed the persistence in processed tomatoes of the main primary and secondary metabolites already reported in fresh fruits, such as essential amino acids, sugar, organic acids, vitamins, fatty acyls, and phytohormones. Moreover, new insight on specific components never traced before in similar finished samples is provided. Bioactive compounds were detected in all samples, such as oligopeptides with ACE-inhibitor activity, ɣ-aminobutyric acid, alkaloids, and polyphenols (flavonoids, coumarins, and cinnamic acids). Many of these compounds have antioxidant activities, proving the relevance of transformed tomatoes as a source of health-promoting compounds for the human diet. Conclusions: A detailed metabolic profile of commercial tomato puree samples was obtained, and a curated database of metabolites was compiled, which can be useful for multiple purposes, for example, authentication, quality, or nutritional assessments. Full article

In recent years, deep learning-based methods for surface defect detection in steel strips have advanced rapidly. Nevertheless, existing approaches still face several challenges in practical applications, such as insufficient dimensionality of feature information, inadequate representation capability for single-modal samples, poor adaptability to few-shot scenarios, and difficulties in cross-domain knowledge transfer. To overcome these limitations, this paper proposes a multi-modal fusion framework based on graph neural networks for few-shot classification and detection of surface defects. The proposed architecture consists of three core components: a multi-modal feature fusion module, a graph neural network module, and a cross-modal transfer learning module. By integrating heterogeneous data modalities—including visual images and textual descriptions—the method facilitates the construction of a more efficient and accurate defect classification and detection model. Experimental evaluations on steel strip surface defect datasets confirm the robustness and effectiveness of the proposed method under small-sample conditions. The results demonstrate that our approach provides a novel and reliable solution for automated quality inspection of surface defects in the steel industry. Full article

Road safety has emerged as a global issue, driven by the rapid rise in vehicle ownership and traffic congestion. Human error, like distraction, drowsiness, and panic, is the leading cause of road accidents. Conventional driver monitoring systems (DMSs) frequently fail to detect these emotional and cognitive states, limiting their potential to prevent accidents. To overcome these challenges, this work proposes a robust deep learning-based DMS framework capable of real-time detection and response to emotion-driven driver behaviors that pose safety risks. The proposed system employs convolutional neural networks (CNNs), specifically the Inception module and a Caffe-based ResNet-10 with a Single Shot Detector (SSD), to achieve efficient, accurate facial detection and classification. The DMS is trained on a comprehensive and diverse dataset from various public and private sources, ensuring robustness across a wide range of emotions and real-world driving scenarios. This approach enables the model to achieve an overall accuracy of 98.6%, an F1 score of 0.979, a precision of 0.980, and a recall of 0.979 across the four emotional states. Compared with existing techniques, the proposed model strikes an effective balance between computational efficiency and complexity, enabling the precise recognition of driving-relevant emotions, making it a practical and high-performing solution for real-world in-car driver monitoring systems. Full article

Time and frequency standards constitute fundamental requirements for diverse applications spanning daily life technologies to advanced scientific research. Among precision time dissemination methods, microwave-clock-based dual comb time transfer has emerged as a promising approach that achieves ultra-precise time interval measurements through linear optical sampling. However, conventional peak detection methodologies employed in such systems exhibit critical limitations: vulnerability to amplitude noise interference and inherent accuracy constraints imposed by analog sampling rates. To address these challenges, we present a novel digital time differential measurement paradigm integrating three key algorithmic innovations: (1) adaptive signal detection and extraction protocols, (2) multi-stage noise suppression processing, and (3) optimized centroid determination techniques. This comprehensive digital processing framework significantly enhances both measurement stability and operational efficiency, demonstrating single-shot temporal resolution at 17.6 fs stability levels. Our method establishes new capabilities for high-precision time-frequency transfer applications requiring robust noise immunity and enhanced sampling dynamics. Full article

As network automation and self-organizing networks (SONs) rapidly evolve, edge devices increasingly demand lightweight, real-time, and high-precision classification algorithms to support critical tasks such as traffic identification, intrusion detection, and fault diagnosis. In recent years, cross-entropy (CE) loss has been widely adopted in deep learning classification tasks due to its computational efficiency and ease of optimization. However, traditional CE methods primarily focus on class separability without explicitly constraining intra-class compactness and inter-class boundaries in the feature space, thereby limiting their generalization performance on complex classification tasks. To address this issue, we propose a novel classification loss framework—Contrastive Geometric Cross-Entropy (CGCE). Without incurring additional computational or memory overhead, CGCE explicitly introduces learnable class representation vectors and constructs the loss function based on the dot-product similarity between features and these class representations, thus explicitly reinforcing geometric constraints in the feature space. This mechanism effectively enhances intra-class compactness and inter-class separability. Theoretical analysis further demonstrates that minimizing the CGCE loss naturally induces clear and measurable geometric class boundaries in the feature space, a desirable property absent from traditional CE methods. Furthermore, CGCE can seamlessly incorporate the prior knowledge of pretrained models, converging rapidly within only a few training epochs (for example, on the CIFAR-10 dataset using the ViT model, a single training epoch is sufficient to reach 99% of the final training accuracy.) Experimental results on both text and image classification tasks show that CGCE achieves accuracy improvements of up to 2% over traditional CE methods, exhibiting stronger generalization capabilities under challenging scenarios such as class imbalance, few-shot learning, and noisy labels. These findings indicate that CGCE has significant potential as a superior alternative to traditional CE methods. Full article

This paper presents HE-DMDeception, a novel adversarial attack network that integrates human visual deception with deep model deception to enhance the security of 3D object detection. Existing patch-based and camouflage methods can mislead deep learning models but struggle to generate visually imperceptible, high-quality textures. Our framework employs a CycleGAN-based camouflage network to generate highly camouflaged background textures, while a dedicated deception module disrupts non-maximum suppression (NMS) and attention mechanisms through optimized constraints that balance attack efficacy and visual fidelity. To overcome the scarcity of annotated vehicle data, an image segmentation module based on the pre-trained Segment Anything (SAM) model is introduced, leveraging a two-stage training strategy combining semi-supervised self-training and supervised fine-tuning. Experimental results show that the minimum P@0.5 values (50%, 55%, 20%, 25%, 25%) were achieved by HE-DMDeception across You Only Look Once version 8 (YOLOv8), Real-Time Detection Transformer (RT-DETR), Fast Region-based Convolutional Neural Network (Faster-RCNN), Single Shot MultiBox Detector (SSD), and MaskRegion-based Convolutional Neural Network (Mask RCNN) detection models, while maintaining high visual consistency with the original camouflage. These findings demonstrate the robustness and practicality of HE-DMDeception, offering new insights into 3D object detection adversarial attacks. Full article

Remote sensing visual question answering (RSVQA) involves interpreting complex geospatial information captured by satellite imagery to answer natural language questions, making it a vital tool for observing and analyzing Earth’s surface without direct contact. Although numerous studies have addressed RSVQA, most have focused primarily on answer accuracy, often overlooking the underlying reasoning capabilities required to interpret spatial and contextual cues in satellite imagery. To address this gap, this study presents a comprehensive evaluation of four large multimodal models (LMMs) as follows: GPT-4o, Grok 3, Gemini 2.5 Pro, and Claude 3.7 Sonnet. We used a curated subset of the EarthVQA dataset consisting of 100 rural images with 29 question–answer pairs each and 100 urban images with 42 pairs each. We developed the following three task-specific frameworks: (1) Zero-GeoVision, which employs zero-shot prompting with problem-specific prompts that elicit direct answers from the pretrained knowledge base without fine-tuning; (2) CoT-GeoReason, which enhances the knowledge base with chain-of-thought prompting, guiding it through explicit steps of feature detection, spatial analysis, and answer synthesis; and (3) Self-GeoSense, which extends this approach by stochastically decoding five independent reasoning chains for each remote sensing question. Rather than merging these chains, it counts the final answers, selects the majority choice, and returns a single complete reasoning chain whose conclusion aligns with that majority. Additionally, we designed the Geo-Judge framework to employ a two-stage evaluation process. In Stage 1, a GPT-4o-mini-based LMM judge assesses reasoning coherence and answer correctness using the input image, task type, reasoning steps, generated model answer, and ground truth. In Stage 2, blinded human experts independently review the LMM’s reasoning and answer, providing unbiased validation through careful reassessment. Focusing on Self-GeoSense with Grok 3, this framework achieves superior performance with 94.69% accuracy in Basic Judging, 93.18% in Basic Counting, 89.42% in Reasoning-Based Judging, 83.29% in Reasoning-Based Counting, 77.64% in Object Situation Analysis, and 65.29% in Comprehensive Analysis, alongside RMSE values of 0.9102 in Basic Counting and 1.0551 in Reasoning-Based Counting. Full article

This work presents a novel time-to-digital converter based on the analog sampling of dual-phase periodic signals generated from a free-running oscillator. A proof-of-concept ASIC, implemented in 130 nm CMOS technology, achieves an average single-shot precision of 0.9 ps-rms for time intervals up to 3 ns, with a best performance of 0.79 ps-rms. It maintains a precision below 3.7 ps-rms for intervals up to 25 ns. The design demonstrates excellent linearity, with a peak-to-peak differential nonlinearity of 0.56 LSB and a peak-to-peak integral nonlinearity of 1.43 LSB. The free-running oscillator is shareable across multiple channels, enabling power consumption of approximately 4.1 mW per channel and efficient area utilization. These features make the design highly suitable for detection systems requiring picosecond-level precision and high channel density, such as silicon pixel sensors, SPADs, LiDARs, and time-correlated single-photon counting systems. Furthermore, the architecture shows strong potential for use in high-count-rate applications, reaching up to 22 Mcps. Full article