Search Results (89)

This study presents a comprehensive comparison of static and SLAM (Simultaneous Localization and Mapping) laser scanners of both new and old generation in a controlled indoor environment of a standard commercial building with long, linear corridors and recesses. The aim was to assess both global and local accuracy, as well as noise characteristics, of each scanner. Methods: A highly accurate static scanner was used to generate a reference point cloud. Five devices were evaluated: two static scanners (Leica RTC 360 and Trimble X7) and three SLAM scanners (GeoSLAM ZEB Horizon RT, Emesent Hovermap ST-X, and FARO Orbis). Accuracy analysis included systematic and random error assessment, axis-specific displacement evaluation, and profile-based local accuracy measurements. Additionally, noise was quantified before and after data smoothing. Static scanners yielded superior accuracies, with the Leica RTC 360 achieving the best performance (absolute accuracy of 1.2 mm). Among SLAM systems, the Emesent Hovermap ST-X and FARO Orbis—both newer-generation devices—demonstrated significant improvements over the older-generation GeoSLAM ZEB Horizon RT. After smoothing, the noise levels of these new-generation SLAM scanners (approx. 2.1–2.2 mm) approached those of static systems. The findings underline the ongoing technological progress in SLAM systems, with the new-generation SLAM scanners becoming increasingly viable alternatives to static scanners, especially when speed, ease of use, and reduced occlusions are prioritized. This makes them well-suited for rapid indoor mapping applications, provided that the slightly lower accuracy is acceptable for the intended use. Full article

Ground-based synthetic aperture radar (GBSAR) has been widely used in the fields of early warning of geologic hazards and deformation monitoring of engineering structures due to its characteristics of high spatial resolution, zero spatial baseline, and short revisit period. However, in the continuous monitoring process of GBSAR, due to the sudden failure of radar equipment, such as power failure, or the influence of alternating work between multiple regions, it often leads to discontinuous data collection, and this problem caused by missing data is collectively called “inspection mode”. The problem of missing data in the inspection mode not only destroys the spatial and temporal continuity of the data but also affects the accuracy of the subsequent deformation analysis. In order to solve this problem, in this paper, we propose a data reconstruction method that combines Sage–Husa Kalman adaptive filtering and the Rauch–Tung–Striebel (RTS) smoothing algorithm. The method is based on the principle of Kalman filtering and solves the problem of “model mismatch” caused by the fixed noise statistics of traditional Kalman filtering by dynamically adjusting the noise covariance to adapt to the non-stationary characteristics of the observed data. Subsequently, the Rauch–Tung–Striebel (RTS) smoothing algorithm is used to process the preliminary filtering results to eliminate the cumulative error during the period of missing data and recover the complete and smooth deformation time series. The experimental and simulation results show that this method successfully restores the spatial and temporal continuity of the inspection data, thus improving the overall accuracy and stability of deformation monitoring. Full article

The growth of Internet of Things (IoT) systems has brought serious security concerns, especially in asymmetrical environments where device capabilities and communication flows vary widely. Many machine-learning-based intrusion detection systems struggle to address noise, uncertainty, and class imbalance. For that reason, intensive data preprocessing procedures were required. These challenges are in real-world data. In this work, we introduce a semi-supervised learning approach that uses entropy-based uncertainty filtering to improve intrusion detection in IoT environments. By dynamically identifying uncertain predictions from tree-based classifiers, we retain only high-confidence results during training. Later, confident samples from the uncertain set are used to retrain the model through a self-training loop. We evaluate this method using three diverse and benchmark datasets named RT-IoT2022, CICIoT2023, and CICIoMT2024, which include up to 34 different attack types. The experimental results reveal that XGBoost and Random Forest outperformed other tree-based models while maintaining their robustness when predicting attacks in the IoT environment. In addition, our proposed model was compared with other models proposed by researchers in the field, and the findings confirmed that our model presented promising results. Full article

Nighttime object detection poses significant challenges due to low illumination, noise, and reduced contrast, which can severely impact the performance of standard detection models. In this paper, we present NF-DETR (Night-Frequency Detection Transformer), a novel framework that leverages frequency domain information to enhance object detection in challenging nighttime environments. Our approach integrates physics-prior enhancement to improve the visibility of objects in low-light conditions, frequency domain feature extraction to capture structural information potentially lost in the spatial domain, and window cross-attention fusion that efficiently combines complementary features while reducing computational complexity, significantly improving detection performance without increasing the parameter count. Extensive experiments on two challenging nighttime detection benchmarks, BDD100K-Night and City-Night3K, demonstrate the effectiveness of our approach. Compared to strong baselines such as YOLOv8-M, YOLOv12-X, and RT-DETRv2-50, NF-DETR-L achieves improvements of up to +3.5% AP@50 and +3.7% AP@50:95 on BDD100K-Night, and +2.7% AP@50 and +1.9% AP@50:95 on City-Night3K, while maintaining competitive inference speeds. Ablation studies confirm that each proposed component contributes positively to detection performance, with their combination yielding the best results. NF-DETR offers a more robust solution for nighttime perception systems in autonomous driving and surveillance applications, effectively addressing the unique challenges of low-light object detection. Full article

Although existing motion transfer methods for bionic robot arms are based on kinematic equivalence or simplified dynamic models, they frequently fail to tackle dynamic compliance and real-time adaptability in complex human-like motions. To address this shortcoming, this study presents a motion transfer method from the human arm to a bionic robot arm based on the hybrid PSO-RF (Particle Swarm Optimization-Random Forest) algorithm to improve joint space mapping accuracy and dynamic compliance. Initially, a high-precision optical motion capture (Mocap) system was utilized to record human arm trajectories, and Kalman filtering and a Rauch–Tung–Striebel (RTS) smoother were applied to reduce noise and phase lag. Subsequently, the joint angles of the human arm were computed through geometric vector analysis. Although geometric vector analysis offers an initial estimation of joint angles, its deterministic framework is subject to error accumulation caused by the occlusion of reflective markers and kinematic singularities. To surmount this limitation, this study designed five action sequences for the establishment of the training database for the PSO-RF model to predict joint angles when performing different actions. Ultimately, an experimental platform was built to validate the motion transfer method, and the experimental verification showed that the system attained high prediction accuracy (R² = 0.932 for the elbow joint angle) and real-time performance with a latency of 0.1097 s. This paper promotes compliant human–robot interaction by dealing with joint-level dynamic transfer challenges, presenting a framework for applications in intelligent manufacturing and rehabilitation robotics. Full article

Blended Synchronous Learning helps teachers and students communicate without geographical restrictions. The effect of communication between the face-to-face end and the remote end was not only affected by the performance of the equipment but also by the acoustic conditions in the classroom. This paper measured the acoustic parameters in the hybrid classrooms and conducted subjective speech intelligibility tests. It was found that for the hybrid classroom with a decentralized sound reinforcement system, the background noise level was high because lots of equipment was needed for synchronous learning. The speech intelligibility scores of the remote end were lower than those at the face-to-face end. Acoustic parameters of reverberation time (RT) and excessive signal-to-noise ratio (SNR) showed a negative correlation with speech intelligibility scores in the remote end. It was recommended that the sound pressure level (SPL) of the sound reinforcement system should not be too high and that appropriate sound absorption treatment be performed. The size of the hybrid classroom should be controlled to prevent the sound that arrived 50 ms after the direct sound from arriving. When SNR was 33 dB(A) for hybrid classrooms, which had a good performance in the face-to-face end with the speech intelligibility scores, T₂₀ should be within 0.8 s to achieve the target value of 83% for SI scores at the remote end. Full article

This study aimed to detect weeds in vineyards throughout the crop cycle using pixel-based classification of RGB imagery captured by unmanned aerial vehicles (UAVs). Five vegetation indices (NGRDI, NDVI, GLI, NDRE, and GNDVI) and three supervised classifiers (SVM, RT, and KNN) were evaluated during four flight campaigns. Classification performance was assessed using precision, recall, and F1-Score, supported by descriptive statistics (mean, standard deviation, and 95% confidence interval), inferential tests (Shapiro–Wilk, ANOVA, and Kruskal–Wallis), and visual map inspection. Statistical analyses, both descriptive and inferential, did not indicate significant differences between classification methods. NGRDI consistently showed strong performance, especially for vine and soil classes, and effectively detected weeds, with F1-Scores above 0.78 in some campaigns, occasionally outperforming the supervised classifiers. GLI displayed variable results and a higher sensitivity to noise, whereas NDVI showed limitations when applied to RGB data, particularly in sparsely vegetated areas. Among the classifiers, the SVM achieved the highest F1-Score for vine (0.9330) and soil (0.9231), whereas KNN produced balanced results and visually coherent maps. RT showed lower accuracy and greater variability, particularly in the weed class. Despite the lack of statistically significant differences, visual analysis favored NGRDI and SVM for generating cleaner classification outputs. Study limitations include lighting variability, reduced spatial coverage owing to low flight altitude, and a lack of spatial context in pixel-based methods. Future research should explore object-based approaches and advanced classifiers (e.g., Random Forest and Convolutional Neural Networks) to enhance robustness and generalization. Overall, RGB-based indices, particularly NGRDI, are cost-effective and reliable tools for weed detection, thereby supporting scalable precision in viticulture. Full article

Despite the pivotal role of filtering and smoothing techniques in the preprocessing of ship maneuvering data for robust identification, persistent challenges in reconciling noise suppression with dynamic fidelity preservation have limited algorithmic advancements in recent decades. We propose an online smoothing method enhanced by the Expectation-Maximization (EM) algorithm framework that effectively extracts high-fidelity dynamic features from raw maneuvering data, thereby enhancing the fidelity of subsequent ship identification systems. Our method effectively addresses the challenges posed by heavy-tailed Student-t distributed noise and parameter uncertainty inherent in ship motion data, demonstrating robust parameter learning capabilities, even when initial ship motion system parameters deviate from real conditions. Through iterative data assimilation, the algorithm adaptively calibrates noise distribution parameters while preserving motion smoothness, achieving superior accuracy in velocity and heading estimation compared to conventional Rauch–Tung–Striebel (RTS) smoothers. By integrating parameter adaptation within the smoothing framework, the proposed method reduces motion prediction errors by 23.6% in irregular sea states, as validated using real ship motion data from autonomous navigation tests. Full article

Defect detection in industrial computed tomography (CT) images remains challenging due to small defect sizes, low contrast, and noise interference. To address these issues, we propose Defect R-CNN, a novel detection framework designed to capture the structural characteristics of defects in CT images. The model incorporates an edge-prior convolutional block (EPCB) that guides to focus on extracting edge information, particularly along defect boundaries, improving both localization and classification. Additionally, we introduce a custom backbone, edge-prior net (EP-Net), to capture features across multiple spatial scales, enhancing the recognition of subtle and complex defect patterns. During inference, the multi-branch structure is consolidated into a single-branch equivalent to accelerate detection without compromising accuracy. Experiments conducted on a CT dataset of nuclear graphite components from a high-temperature gas-cooled reactor (HTGR) demonstrate that Defect R-CNN achieves average precision (AP) exceeding 0.9 for all defect types. Moreover, the model attains mean average precision (mAP) scores of 0.983 for bounding boxes (mAP-bbox) and 0.956 for segmentation masks (mAP-segm), surpassing established methods such as Faster R-CNN, Mask R-CNN, Efficient Net, RT-DETR, and YOLOv11. The inference speed reaches 76.2 frames per second (FPS), representing an optimal balance between accuracy and efficiency. This study demonstrates that Defect R-CNN offers a robust and reliable approach for industrial scenarios that require high-precision and real-time defect detection. Full article

Wind energy plays a crucial role in global sustainable development, and accurately estimating the number and spatial distribution of wind turbines is crucial for strategic planning and energy allocation. To address the critical need for wind turbine detection and spatial distribution analysis, this study develops YOLOv5_CDB, an enhanced detection framework based on the YOLOv5 model. The proposed method incorporates two key components: the Convolutional Block Attention Mechanism (CBAM) to improve feature representation and the Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm for spatial density clustering. The method is applied to 2 m resolution World Imagery data. It detects both tubular and lattice wind turbines by analyzing key features, including turbine towers and shadows. The YOLOv5_CDB demonstrates a substantial enhancement in performance when compared with the YOLOv5s. The F1-score shows an increase of 1.39%, and the mean average precision (mAP) exhibits a 1.5% improvement. Meanwhile, the precision (P) and recall (R) values are recorded at 95.97% and 91.18%, respectively. Furthermore, YOLOv5_CDB evinces consistent performance advantages, outperforming state-of-the-art models including YOLOv8s, YOLOv12s, and RT-DETR by 1.84%, 3.98%, and 1.77% in terms of F1-score and by 3.7%, 4.5%, and 3.0% in terms of mAP, respectively. The YOLOv5_CDB model has been demonstrated to show superior performance in the global wind turbine detection domain, thereby providing a foundation for the management of wind farms and the development of sustainable energy. Full article

The leakage signal of the hydraulic valve is a weak, nonlinear, and non-periodic signal that is easily overpowered by background noise from the surroundings. To address this issue, the Search and Rescue Team (SaRT) algorithm was introduced to adaptive coupled stochastic resonance, and a new signal-enhancement method based on SaRT for coupled multi-stable stochastic resonance (CMSR) was proposed for enhancing valve-leakage vibration signals. Initially, the method employs the rescaling technique to preprocess the signal, thereby transforming the fault signal into a small-parameter signal. Subsequently, the mutual correlation gain is utilized as an adaptive measure function of the SaRT algorithm to optimize the parameters of the coupled multi-stable stochastic resonance system. Ultimately, the output signal is solved by the fourth-order Runge–Kutta method. This study validated the method using sinusoidal signals and leakage signals of the check valve. The results demonstrate that all CMSR parameters require optimization. Furthermore, the noise reduction was effective for three different leakage signals of faulty check valves, in which the highest in the number of interrelationships increased by 6.9569 times and the highest amplitude ratio of the peak frequency increased by 11.7004 times. The data quality was significantly improved. Full article

In complex underground parking scenarios, non-line-of-sight (NLOS) obstructions significantly impede positioning signals, presenting substantial challenges for accurate vehicle localization. While traditional positioning approaches primarily focus on mitigating NLOS effects to enhance accuracy, this research adopts an alternative perspective by leveraging NLOS propagation as valuable information, enabling precise positioning in NLOS-dominated environments. We introduce an innovative NLOS positioning framework based on the generalized source (GS) technique, which employs ray-tracing (RT) to transform NLOS paths into equivalent line-of-sight (LOS) paths. A novel GS filtering and weighting strategy to establish initial weights for the nonlinear equation system. To combat significant NLOS noise interference, a robust iterative reweighted least squares (W-IRLS) method synergizes initial weights with optimal position estimation. Integrating ultra-wideband (UWB) delay and angular measurements, four distinct localization modes based on W-IRLS are developed: angle-of-arrival (AOA), time-of-arrival (TOA), AOA/TOA hybrid, and AOA/time-difference-of-arrival (TDOA) hybrid. The comprehensive experimental and simulation results validate the exceptional effectiveness and robustness of the proposed NLOS positioning framework, demonstrating positioning accuracy up to 0.14 m in specific scenarios. This research not only advances the state of the art in NLOS positioning but also establishes a robust foundation for high-precision localization in challenging environments. Full article

Compared with traditional single-view and radiometric-only observations, multiangle polarimetric observations of water-leaving radiation play a crucial role in enhancing the retrieval of ocean constituents and aerosol microphysical properties. In this study, the impacts of uncertainties in the degree of polarization (DOP) of water-leaving radiance (L_w) on the retrieval of oceanic constituents and inherent optical properties (IOPs) were investigated via global radiative transfer (RT) simulations and the fully connected U-Net (FCUN) model. The uncertainties in the retrieval of oceanic constituents and IOPs were further investigated with various sensor azimuth angles. The results indicated that the global mean absolute percentage errors (MAPEs) for differing oceanic constituents and IOPs significantly decreased as the number of observation angles increased. Taking the retrieval of Chla as an example, the global MAPEs between the FCUN predictions and RT simulation inputs for Chla concentrations under differing observation angles were 7.41%, 3.76%, 2.70%, 2.44%, 2.62%, and 1.82%. Moreover, the MAPEs at sensor azimuth angles of 0° and 30° were significantly lower than those at other azimuth angles for the single-view observations. As the number of observation angles increased, the variation in MAPEs with the sensor azimuth angle gradually weakened. Furthermore, the impact of errors in the L_w DOP on the retrieval uncertainties decreased as the number of observation angles increased, and the global MAPEs of Chla after adding the various random instrument noises were 46.56% (46.91%), 6.59% (7.21%), 5.21% (5.79%), 4.72% (4.98%), 3.99% (4.52%), and 3.64% (4.03%). Overall, the multiangle polarimetric observations can suppress or balance the impact of uncertainties in the L_w DOP on the retrieval of oceanic constituents and IOPs. Full article

In this research, a multi-channel target speech enhancement scheme is proposed that is based on deep learning (DL) architecture and assisted by multi-source tracking using a labeled random finite set (RFS) framework. A neural network based on minimum variance distortionless response (MVDR) beamformer is considered as the beamformer of choice, where a residual dense convolutional graph-U-Net is applied in a generative adversarial network (GAN) setting to model the beamformer for target speech enhancement under reverberant conditions involving multiple moving speech sources. The input dataset for this neural architecture is constructed by applying multi-source tracking using multi-sensor generalized labeled multi-Bernoulli (MS-GLMB) filtering, which belongs to the labeled RFS framework, to obtain estimations of the sources’ positions and the associated labels (corresponding to each source) at each time frame with high accuracy under the effect of undesirable factors like reverberation and background noise. The tracked sources’ positions and associated labels help to correctly discriminate the target source from the interferers across all time frames and generate time–frequency (T-F) masks corresponding to the target source from the output of a time-varying, minimum variance distortionless response (MVDR) beamformer. These T-F masks constitute the target label set used to train the proposed deep neural architecture to perform target speech enhancement. The exploitation of MS-GLMB filtering and a time-varying MVDR beamformer help in providing the spatial information of the sources, in addition to the spectral information, within the neural speech enhancement framework during the training phase. Moreover, the application of the GAN framework takes advantage of adversarial optimization as an alternative to maximum likelihood (ML)-based frameworks, which further boosts the performance of target speech enhancement under reverberant conditions. The computer simulations demonstrate that the proposed approach leads to better target speech enhancement performance compared with existing state-of-the-art DL-based methodologies which do not incorporate the labeled RFS-based approach, something which is evident from the 75% ESTOI and PESQ of 2.70 achieved by the proposed approach as compared with the 46.74% ESTOI and PESQ of 1.84 achieved by Mask-MVDR with self-attention mechanism at a reverberation time (RT60) of 550 ms. Full article

This article proposes a force rebalance control scheme based on a mode-localized resonant accelerometer (ML-RXL), which is applied to address the limited measurement range problem of the ML-RXL. For the first time, an empirical response model of the weakly coupling resonators for the amplitude ratio output is established. Based on this, this paper builds an overall model of the force rebalance control system to analyze the sensitivity characteristics by simulations, which demonstrates that the scheme can effectively broaden the linear measurement range. It is demonstrated that the sensor exhibits a highly linear output within a measurement range of ±1 g, with a sensitivity of the feedback-control voltage output measured at 2.94 V/g. The measurement range is expanded by at least 6.7 times. Moreover, the results show that the minimum input-referred acceleration noise density of the sensor for the force rebalance control scheme is 3.29 μg/rtHz, and that the best bias instability is optimized to 5.34 μg with an integral time of 0.64 s. Full article