High-resolution time-of-flight mass spectrometry (TOF-MS) imposes rigorous performance demands on high-voltage pulse generators (HVPGs): extremely short edges for rapid electric field establishment, adjustable widths for ion mass range control, and stable amplitudes for precisely imparting ion kinetic energy. Existing pulse generators struggle to meet these multidimensional demands simultaneously. This article proposes a hybrid HVPG topology that utilizes a Marx-based sharpening circuit to augment conventional half-bridge pulsers significantly. A mechanistic model is established to clarify parameter selection for optimized output characteristics. A developed prototype achieved 4.5 ns falling edges in laboratory tests, with widths adjustable from 2 mu s to 100% duty cycle, a repetition frequency of 0-2 kHz, and amplitudes spanning 600-800 V. In-system validation under actual TOF-MS loads demonstrated 6 ns edges, providing a 59.3% mass resolution improvement over standalone half-bridge pulsers. The results indicate that the HVPG approach effectively satisfies rigorous TOF-MS requirements.

This paper presents UM3D, an end-to-end unsupervised domain adaptation framework for monocular 3D object detection. Monocular 3D object detection is appealing due to its low cost, yet it suffers from limited depth cues and poor cross-domain generalization when labeled data are scarce. Existing Pseudo-LiDAR methods require supervised training and propagate depth estimation errors to downstream detection, while current unsupervised domain adaptation (UDA) approaches exploit only a single modality and lack effective pseudo-label quality control. UM3D addresses these limitations through two key designs: (1) a quality-aware pseudo-label generation strategy with object-level random scaling and a memory bank refinement mechanism; and (2) an end-to-end differentiable pipeline that integrates multimodal fusion of image and Pseudo-LiDAR features with a multi-network consistency loss, which jointly optimizes depth estimation and 3D detection via backpropagation. Notably, the entire pipeline requires only a single monocular camera at inference; the Pseudo-LiDAR representation is generated internally from the same image, and thus the multimodal fusion integrates image and Pseudo-LiDAR features without requiring additional sensors. Extensive experiments across KITTI, nuScenes, Waymo, and Lyft demonstrate that UM3D generally outperforms existing UDA methods. In particular, a 19.30% relative APBEV improvement is achieved under easy conditions through end-to-end joint training compared to independent depth estimation, and up to 76.81% of the domain gap is closed on the WOD → KITTI benchmark.

Reliable acquisition of underwater surface electromyography (sEMG) is the cornerstone of aquatic health monitoring, yet it faces a critical challenge: the inevitable swelling of hydrogel electrodes disrupts the skin-electrode interface, leading to severe signal degradation. To address this dilemma, we propose a "Dual-Locking" densification strategy synergizing internal ionic locking (electrostatic interaction) and solvent-regulated physical crystallization to construct a robust polyphenol-based hydrogel (PTDG). This strategy effectively suppresses the swelling ratio (<3% in water; <9% in seawater) while preserving high ionic conductivity (1.84 S & centerdot;m(-1)) and strong wet adhesion (9.33 kPa), successfully decoupling the trade-off between stability and electrochemical performance. Consequently, the PTDG electrode maintains a stable contact interface with minimal impedance drift, enabling high-fidelity sEMG recording with a signal-to-noise ratio of 36.47 dB even after 60 min of underwater operation. Validated by a rigorous session-wise deep learning evaluation (achieving a gesture recognition accuracy of 97.59% and a Macro-F1 score of 97.09%) and precise muscle fatigue monitoring, this work presents a versatile solution for reliable underwater biopotential acquisition.

Detecting drivable areas is a fundamental task in autonomous driving systems. Although semantic segmentation networks have demonstrated strong performance in segmenting drivable regions, two key challenges persist. First, acquiring sufficient contextual information in complex road scenarios remains difficult, often leading to segmentation errors. Second, the coarseness of extracted features may degrade accuracy even when texture information is available in RGB images. To address these issues, we propose an enhanced DeepLabv3+ algorithm called Split Convolution Selective Attention Network (SCSANet), which incorporates the Adaptive Kernel (AK) and Split Convolution Attention (SCA) modules. AK adaptively adjusts the receptive field to accommodate varying road scenarios, while SCA improves boundary clarity by enhancing channel interaction. In addition, we employ surface normals to provide complementary geometric information, thereby strengthening the ability of the network to recognize drivable areas. To compensate for the lack of publicly available datasets for closed or semi-closed scenarios, we introduce XMUROAD, a new dataset of binocular disparity images. Experiments on the XMUROAD dataset demonstrate that the proposed architectural improvements yield an mIoU gain of 1.63% under the same RGB input, and the full pipeline with surface normal input achieves improvements of 1.55% to 2.59% in mF1 and 2.94% to 4.83% in mIoU over state-of-the-art methods. Experiments on the KITTI dataset further verify the generalization capability of SCSANet, with improvements of 1.58% in mF1 and 2.88% in mIoU over state-of-the-art methods. The proposed method provides a practical approach for accurate drivable area detection in closed and semi-closed mobile-robot scenarios.

Road potholes pose a considerable threat to mobile robots, which are generally less stable than conventional vehicles and may become trapped or overturned when traversing damaged road surfaces. Accurate semantic segmentation of road potholes is therefore essential for safe and reliable robot navigation. To address this requirement, multimodal fusion methods using RGB (Red, Green, Blue) and disparity images have been developed for pothole detection. Nevertheless, these methods still face challenges in detecting small potholes and delineating their boundaries precisely. To overcome these limitations, we propose a novel multimodal fusion network for road-pothole semantic segmentation. Specifically, we design a feature fusion module that integrates global context and local details to fully exploit the complementary information provided by RGB and disparity images. This design improves multimodal feature interaction and enhances boundary segmentation accuracy. Furthermore, we develop three feature attention fusion modules by incorporating multiple complementary attention mechanisms into the fusion module. These modules improve small-pothole detection by focusing on informative features, emphasizing target regions, and reducing information loss. We evaluate the proposed network on a small-pothole subset of Pothole-600 under identical hardware settings and backbone configurations for all experimental models. On the small-pothole subset of Pothole-600, FAFMNet achieves 90.22% mPre, 92.32% mRec, 98.73% mAcc, 91.26% mF1, and 83.93% mIoU, outperforming the state-of-the-art method by 1.87 percentage points in mF1 and 3.12 percentage points in mIoU. A paired statistical test over three independent runs further confirms that the improvement over the baseline is statistically significant (p<0.05).

Purpose - This paper aims to propose a method for automated operation of electrical switchgear based on the integration of visual positioning guidance and force perception. The method includes a visual algorithm for locating targets and guiding the robotic arm, and a visual positioning and guidance error compensation strategy based on force sensing. Design/methodology/approach - This paper presents an algorithm for the recognition and localization of operational targets in electric power switchgear cabinets, employing a low-cost RGB-D camera. The algorithm applies Deep Convolutional Neural Networks (DCNN) for detection purposes and achieves stable and accurate identification and localization of operational targets through the integration of RGB images and depth data. Furthermore, a force perception prediction model based on a Backpropagation (BP) neural network is established in this paper to compensate for and correct errors in vision-based localization guidance, thereby facilitating automated operation of electrical switchgear. Findings - In the experiments, the relative pose between the operating platform and the electrical switchgear was altered to simulate navigation and positioning deviations, and guidance and positioning operations were tested using the methods outlined in this paper. The results indicate that within the scope of navigation and positioning deviation, the guided positioning accuracy of the proposed method in this paper is 2.6 mm and 0.5 degrees. Furthermore, a practical application was conducted in the Power distribution room, and the results demonstrate that this method can achieve automated operation of electrical switchgear. Research limitations/implications - Currently, the method presented in this paper can be practically applied to the automated operation of specific types of electrical switchgear. In the future, further optimization of the algorithmic process will be pursued to enhance the universality of the algorithm, thereby meeting the automated operation requirements for various types of electrical switchgear. Practical implications - The method proposed in this paper can replace manual operators to perform high-risk switchgear operations and promote intelligent operation and maintenance in the power industry. Additionally, the camera and force sensor used in this method are low-cost and small, making it economically and practically beneficial. Originality/value - This paper proposes an automated operation method for electrical switchgear based on the integration of visual recognition and force perception. This method is applicable to the automated inspection of electric power switchgears, achieving a high operational accuracy of 2.6 mm and 0.5 degrees.

High-spatial-resolution mass spectrometry imaging (MSI) visualizes molecular distributions in tissues and cells. However, achieving higher spatial resolution typically necessitates smaller pixel dimensions and an increased number of pixels, leading to longer data acquisition times and diminished analytical throughput. Although deep learning approaches have demonstrated significant potential in MSI, they typically require large training data sets or paired images, which are often unavailable. Herein, we propose the reference-based super-resolution for mass spectrometry imaging (RSR-MSI) method, with optical microscopy images as reference frames to extract abundant texture information. By integrating this with ion intensity data from the original MS images, we develop an image-specific super-resolution network. Employing solely a single low-resolution MS image coupled with a reference optical image, we successfully reconstruct high-resolution MS images for biological tissues and single cells, producing results with rich chemical and textural details. This approach significantly decreases the routine pixel-by-pixel scanning time by an order of magnitude while achieving high spatial resolution using existing mass spectrometry instruments without any customized modifications. Overall, our work introduces and validates the application of image super-resolution methods within the realm of single-cell MSI at subcellular resolution, paving the way for the development of high-spatial-resolution and high-throughput MSI for cellular biology research.

Depth estimation is an essential component of computer vision applications for environment perception, 3D reconstruction and scene understanding. Among the available methods, self-supervised monocular depth estimation is noteworthy for its cost-effectiveness, ease of installation and data accessibility. However, there are two challenges with current methods. Firstly, the scale factor of self-supervised monocular depth estimation is uncertain, which poses significant difficulties for practical applications. Secondly, the depth prediction accuracy for high-resolution images is still unsatisfactory, resulting in low utilization of computational resources. We propose a novel solution to address these challenges with three specific contributions. Firstly, an interleaved depth network skip-connection structure and a new depth network decoder are proposed to improve the depth prediction accuracy for high-resolution images. Secondly, a data vertical splicing module is suggested as a data enhancement method to obtain more non-vertical features and improve model generalization. Lastly, a scale recovery module is proposed to recover the accurate absolute depth without additional sensors, which solves the issue of uncertainty in the scale factor. The experimental results demonstrate that the proposed framework significantly improves the prediction accuracy of high-resolution images. In particular, the novel network structure and data vertical splicing module contribute significantly to this improvement. Moreover, in a scenario where the camera height is fixed and the ground is flat, the effect of scale recovery module is comparable to that achieved by using ground truth. Overall, the RSANet framework offers a promising solution to solve the existing challenges in self-supervised monocular depth estimation.

Chitosan, as a hydrophilic, renewable, biodegradable, and cost-effective natural polymer, holds great promise in humidity sensor applications. This study proposes a green high-performance capacitive humidity sensor based on chitosan (CS)-sodium chloride (NaCl) composites. The CS-NaCl films were fabricated through a simple drip coating method, and their surface morphology and structure were characterized using SEM, XRD, FTIR and XPS techniques. The pure chitosan humidity sensor demonstrates an impressive response of 1180.17 % with excellent linearity. The introduction of NaCl markedly boosted the sensor response, with CS-50 wt% NaCl composites exhibiting a 72-fold increase (86, 492.84 %) in capacitance response compared to pure CS films. The contact angle test demonstrated that the chitosan film exhibited good hydrophilicity, which was further enhanced by the addition of NaCl. XRD and FTIR characterization confirmed that NaCl altered the structural properties of chitosan, disrupted hydrogen bonds, and consequently increased the mobility of polarizing groups and the dielectric constant. A water adsorption model was developed to elucidate the humidity sensing mechanism of the CS-NaCl sensor. The CS-NaCl composite material presents advantages such as easy preparation, high sensitivity, ecofriendliness, and promising practical applications in humidity sensing.

High sensitivity is crucial for anisotropic magnetoresistive (AMR) sensors in industrial applications. In this paper, a high- sensitive AMR sensor based on magnetoresistive thin films with Ta/NiFe/Ta/Al four-layer structure is proposed and fabricated. Firstly, the structural parameters were optimized by finite element analysis. Secondly, thin film samples and AMR sensors were prepared. Through the analysis and characterization of reluctance change rate, hysteresis loop, x-ray diffraction and surface morphology and structure, the process parameters were optimized. Finally, the sensor was connected to the designed external circuit, and its technical parameters were tested in a magnetic field test system. The results show that the prepared AMR sensor performs well. It has a high sensitivity of 1.27 mV/V/Oe, a low bridge offset voltage of +/- 1.64 mV V-1, and a low temperature coefficient of sensitivity of -0.102%/degrees C. The results contribute to the future development of AMR magnetic field sensor chips.

AbstractEpidermal electronics that can monitor physiological signals such as surface electromyogram (sEMG) signals attract widespread attentions in personalized healthcare, human–machine interfaces (HMI) and virtual/augmented reality (AR/VR). However, conventional electromyographic electrodes suffer from skin discomfort, susceptibility to motion artifact interference, and short service lifetime. Here, an organohydrogel‐based sEMG electrode endows with high conductivity, low modulus and long‐term stability is developed by doping partially reduced graphene oxide (pRGO) into highly cross‐linked organohydrogel network. The as‐fabricated polyacrylamide/sodium alginate/tannic acid/partially reduced graphene oxide (PAM/SA/TA/pRGO) organohydrogel possesses farewell conductivity (4.22 S m−1) while preserving tissue‐like compliance (Young's modulus ≈32 KPa), excellent stretchability (≈600%), high adhesion as well as superior anti‐drying properties. In addition, a stretchable sEMG electrode for long‐term reliable service is fabricated via immobilizing the organohydrogel electrodes onto a flexible very high bond (VHB) substrate. As a result, the integrated electrodes show high signal‐to‐noise ratio (SNR) (35.15 db) comparable to that of the commercial electrodes. Furthermore, with assistance of deep learning, the proposed sEMG electrodes obtain high identification accuracy of 97.11% in distinguishing sophisticated gestures. This system can be further exploited for real‐time tele‐operations and offers broad prospects in human–machine immersive interactive application.

In recent years, with the increasing complexity of software systems, logs have become crucial for system maintenance. Log-based anomaly detection plays a vital role in automatically detecting system anomalies through log analysis. However, current log-based anomaly detection approaches encounter significant practical challenges. Supervised methods often require a large amount of manually labeled training data, which can be time-consuming and costly to obtain. On the other hand, unsupervised and semi-supervised approaches may suffer from subpar performance, as they do not leverage historical anomalies to improve their detection capabilities. These challenges underscore the necessity for the development of more efficient and effective log-based anomaly detection methods. Database anomaly access detection is critical for ensuring the stability and security of database systems. We present a survey of existing log anomaly detection models and propose a novel approach, Template-Parsed Log Anomaly Detection (TPLAD) model, for automated anomaly detection in massive database log files. The proposed model combines the original log template with template parsing using code and text semantic representation. Experimental results demonstrate the effectiveness of the proposed approach in detecting abnormal database access patterns, including runtime errors, unauthorized access, and data leaks. The findings indicate that TPLAD model shows promise in enhancing database security and stability in business systems.

Pointer meter is widely utilized in the fields of modern industry. Nowadays, intelligent inspection robots are gradually employed in place of labors for inspection task. Pointer meter recognition is one of the most important tasks of inspection robots. This article presents a lightweight pointer meter recognition algorithm, which is suitable for deployment on inspection robots. First, pointer meter is identified by pruned YOLOv5. Afterwards, the dial and the pointer are segmented through improved Deeplabv3+ in which the JPU and depthwise separable convolutions are utilized in lieu of dilated convolutions. After perspective transform and central-line extraction of pointer, the reading of the pointer meter can be determined using the angle method. Experimental results verify the FPS of pruned YOLOv5 improves 13.18% compared to original YOLOv5 and the FPS of improved Deeplabv3+ improves 45.74% compared to original Deeplabv3+. Additionally, the reading accuracy of the algorithm is 98.40% and average fiducial error is 0.32%, which indicate good accuracy. The relative standard deviation of reading is less than 1.4%, which indicates good stability of proposed algorithm. This study proposes a lightweight and accurate pointer meter recognition algorithm based on improved Deeplabv3+, the algorithm is ported to NVIDIA Jetson TX2 NX to verify its stability and accuracy.

Electric power artificial intelligence has rapidly advanced in recent years, encompassing safety detection, assistant decision-making, and optimal scheduling. With the rise of Large Language Models (LLMs), knowledge-based AI is becoming increasingly prevalent across various domains. However, in the field of electric power, most of the knowledge-based AI is centered on Knowledge Graph (KG) techniques, while less research has been done on power LLMs. In this paper, we are inspired by Self-Consistency (SC) and propose a Self-Consistency, Extraction and Rectify framework-SCER, for the usage of KG-enhanced LLM in power operations and maintenance (O&M) question answering scenarios. Specifically, we transfer the SC from the general-purpose domain into the power domain and replace the original model with a Chinese sentence representation model to make it more localized. We design an Extract Mechanism to generate evidence chains through multiple random walks on the POMKG and a Rectify Mechanism to correct the score of the generated rationales. Extensive experiments and specific case studies on the POMQA dataset demonstrate the effectiveness of our proposed SCER for SC transfer and improvement in the power field.

Aiming at solving the issue of blurred images and difficult recognition of digital meters encountered by inspection robots in the inspection process, this paper proposes a deep-learning-based method for blurred image restoration and LED digital identification. Firstly, fast Fourier transform (FFT) is used to perform blur detection on the acquired images. Then, the blurred images are recovered using spatial-attention-improved adversarial neural networks. Finally, the digital meter region is extracted using the polygon-YOLOv5 model and corrected via perspective transformation. The digits in the image are extracted using the YOLOv5s model, and then recognized by the CRNN for digit recognition. It is experimentally verified that the improved adversarial neural network in this paper achieves 26.562 in the PSNR metric and 0.861 in the SSIM metric. The missing rate of the digital meter reading method proposed in the paper is only 1% and the accuracy rate is 98%. The method proposed in this paper effectively overcomes the image blurring problem caused by the detection robot during the detection process. This method solves the problems of inaccurate positioning and low digital recognition accuracy of LED digital meters in complex and changeable environments, and provides a new method for reading digital meters.

We present a camera-based human body parameters measurement approach and develop a human postural assessment system. The approach combines the conventional contact measurement method and the non-contact measurement method to overcome some shortcomings in terms of time, expense, and professionalism in early methods. The entire measurement system consists of a computer, a high-definition camera, and the sticky points that are applied to the participant's body before the measurement. The camera captures the triple view image of human body. Then, the human body outline and the joint points of the human skeleton are extracted to locate the bone feature points. Finally, measurements and extractions of the human parameters are made. Experimental results demonstrate that the global postural assessment system provides quantitative guidance for human postural evaluation, and it completely changes how human postural is evaluated. The postural assessment system is significant for early diagnosis of diseases and medical rehabilitation treatment.

Under outdoor inspection environments,the images captured by inspection robots were easily affected by various factors,resulting in low accuracy and poor robustness of pointer meter recognition.To address this issue,a pointer meter recogni-tion algorithm based on improved PSPNet was proposed.Firstly,YOLOv5 was utilized to identify the meter area.Then improved PSPNet was utilized to extract the pointer and scale area.After perspective transformation,the center-line of pointer was extrac-ted.Finally,the meter reading was calculated via angle method.The experimental results show that the average relative error of pointer meter reading is 1.28%and the average citation error is 0.68%under various complex outdoor environments.The average processing time of each picture is 1.28 seconds.It means our method is of great accuracy and stability,providing an effective pointer meter recognition algorithm for outdoor inspection robots.

In the recognition of distracted driving behaviour, traditional manual feature extraction is subjective and complex; single deep convolutional network also has problems such as insufficient generalisation performance and stability. To solve the above problems, this paper proposes a distracted driving behaviour recognition method based on transfer learning and model fusion. First, based on the transfer learning method, the deep convolutional neural network models ResNet18 and ResNet34 are used to extract the features of some images, respectively. Furthermore, the pre-trained model is fine-tuned to obtain four deep convolutional neural network models. Finally, the four network models are fused by stacking method, using 5-fold cross-validation method to reduce over-fitting. Experimental results show that the recognition accuracy of distracted driving behaviour after model fusion reaches 95.47%. The fusion model has higher model generalisation performance and recognition accuracy, which can provide certain technical support for the research of distracted driving behaviour recognition.

To solve the problem of inflexibility of offline hand–eye calibration in “eye-in-hand” modes, an online hand–eye calibration method based on the ChArUco board is proposed in this paper. Firstly, a hand–eye calibration model based on the ChArUco board is established, by analyzing the mathematical model of hand–eye calibration, and the image features of the ChArUco board. According to the advantages of the ChArUco board, with both the checkerboard and the ArUco marker, an online hand–eye calibration algorithm based on the ChArUco board is designed. Then, the online hand–eye calibration algorithm, based on the ChArUco board, is used to realize the dynamic adjustment of the hand–eye position relationship. Finally, the hand–eye calibration experiment is carried out to verify the accuracy of the hand–eye calibration based on the ChArUco board. The robustness and accuracy of the proposed method are verified by online hand–eye calibration experiments. The experimental results show that the accuracy of the online hand–eye calibration method proposed in this paper is between 0.4 mm and 0.6 mm, which is almost the same as the offline hand–eye calibration accuracy. The method in this paper utilizes the advantages of the ChArUco board to realize online hand–eye calibration, which improves the flexibility and robustness of hand–eye calibration.

Currently, pointer meters in substations are read manually, which causes a waste of human resources. To realize automatic reading of such meters, an accurate method for recognizing and reading of pointer meters was proposed for the inspection robot in substation. The new method consists of three steps. Step 1: the image is preprocessed with grayscale, noise reduction and piecewise linear grayscale transformation. Step 2: the Gaussian scale space is used to enhance the scale invariace of the ORB algorithm. Meanwhile, the RANSAC algorithm is used to screen the matching point pairs, which improves the accuracy of the feature pointer matching. Step 3: the Hough transform algorithm is utilized to fit the pointer. Finally, the reading result is obtained according to the angle method and the prior information of the pointer meter. The proposed approach can recognize and read the meter accurately. It has high reliability and engineering application value.