Search

In aeroengine maintenance, endoscopic imaging serves as a crucial tool for detecting blade defects and evolves toward intelligence driven by computer vision technology. Currently, supervised-learning-based defect segmentation methods mainly rely on extensive pixel-level annotations, making it laborious and time consuming. This article shifts focus to the abundant unlabeled data in real-world scenarios and introduces an innovative semisupervised defect segmentation method termed SAIT. Within this framework, three parallel self-supervised mechanisms are adeptly integrated with a semisupervised framework, aiming to bolster defect semantic segmentation with limited labeled samples. In the initial phase, by leveraging the capability of the vision transformer to dissect images into patches, four stochastic distortions are seamlessly infused into the patch sequence. Subsequently, three self-supervised tasks from image level to pixel level are achieved through a customized joint objective function paired with a tailored backbone network. In the second phase, SAIT undergoes pixel-level fine-tuning via the proposed class-centric loss, mitigating class imbalances in limited sample sizes and enhancing initial training. Experiments on a proprietary dataset demonstrate that SAIT achieved 78.42% and 86.70% in mean intersection over union and mean pixel accuracy metrics, respectively, with 25% labeled data, significantly improving the performance of existing semisupervised defect segmentation techniques. Meanwhile, experiments on the open-source dataset further indicate that SAIT holds promise for application in other industrial sectors beyond aeroengine inspection.

In remote sensing image analyses, the extraction of lake water bodies has been emphasized owing to its pivotal role in interpreting aquatic ecosystems, assessing hydrological trends, and detecting environmental changes. Although deep-learning-based techniques have been effectively deployed for this task, several challenges persist, including mis-segmentation of low-contrast regions, insufficient delineation of fuzzy boundaries, and over-segmentation in micro-regions. To address these issues, an innovative, end-to-end segmentation framework is proposed in this study. This framework ingeniously integrates the geodesic active contour (GAC) model with the vision transformer (ViT) architecture, thereby offering a robust solution for fain-gained lake water body extraction. Specifically, the conventional GAC model is first revisited and reformulated. Subsequently, the differentiable spectral clustering (DSC) module is designed for automatic contour initialization and lifts the dependence on hard threshold binarization. Finally, an iterative learning strategy is introduced to integrate the aforementioned models within the ViT architecture. This strategy serves to constrain the geometric characteristics, while also permitting dynamic adjustment of evolution parameters, facilitating enhanced refinement of spatial details and accuracy in segmentation results. Comprehensive experiments conducted on the newly constructed Global Lake View (GLV) dataset and a public benchmark demonstrated that the proposed method yields precise lake boundaries and achieves state-of-the-art performance.

With the rapid development of the aviation industry, videoscope inspection of aeroengines has become crucial for ensuring aircraft flight safety. Recently, deep learning, particularly convolutional neural networks (CNNs), have shown remarkable efficacy in videoscope inspection tasks. However, these methods usually require large-scale training labels with accurate annotation. In videoscope images, defects are often present at the micrometer level and require manual labeling by professional inspection personnel, posing further challenges to model training. To address these issues, weakly supervised deep level set (WDLS), a fully automatic framework for aeroengine defect segmentation, was proposed in this work. WDLS employs a multibranch structure and an iterative learning strategy to combine a high-performance CNN architecture with level set evolution. First, the similarity-guided region detector was designed to generate class-specific segmentation proposals and initialize the level set function. Second, the adaptive local parameter prediction algorithm was proposed to integrate local priors and constraints into the energy function and optimize the iterative process. Finally, a self-supervised objective function named convexified level set was introduced to represent an improved level set formulation and obtain the final segmentation. The proposed framework is continuously differentiable and unified. Thus, it can be seamlessly embedded in any CNN without postprocessing. Furthermore, the method was evaluated on a new aeroengine dataset Turbo19 and a benchmark dataset. Experimental results demonstrated that the proposed framework meets real-time requirements and achieves better performance than state-of-the-art methods.

In this paper, we systematically designed vdW heterostructures by stacking the GaN and Janus-WSSe monolayer and chose the structure that has potential to be used as photocatalysts of water splitting for further investigation. By applying a biaxial strain, the electronic, optical and photocatalytic properties of this heterostructure were investigated to explore whether it is suitable for photocatalyst of water splitting. Results show the heterostructure show a type II band alignment under a − 2% strain, and the band gap is 2.26 eV which is fit to absorb solar light. The band edge of it straddles the water redox potentials at pH = 7. And the absorption coefficient in the visible and UV light region is large, especially for ultraviolet light. In a word, this GaN/Janus-WSSe vdW heterostructure under a − 2% biaxial external strain is a potential solar photocatalyst for water splitting at pH = 7.

Deep supervised learning-based fault diagnosis methods require a large amount of labeled data, which frequently contradicts the typical engineering scenario, in which numerous samples are available but only a small portion is labeled. To conduct a fault diagnosis in this case, unsupervised and semi-supervised representation learning has recently been proposed. However, most of these methods have been designed without much consideration for downstream classification tasks and, thus, are insufficiently relevant for providing sufficient targeted assistance. Therefore, this study proposes a semi-self-supervised method (S3M) that consists of two learning stages. In the self-supervised learning (SSL) stage, two pretext tasks are implemented. Using two auxiliary feature extractors and classifiers, the global and local features of the unlabeled samples can be learned by the feature extractor. In the supervised learning stage, the feature extractor parameters are frozen, and the classifier is trained with the limited number of labeled samples available. The performance of the proposed method was verified using the Case Western Reserve University (CWRU) open dataset and a self-built experimental dataset. Experiments on these two datasets demonstrated that the proposed method outperformed existing diagnosis methods for a few labeled samples.

Sleep-stage and apnea–hypopnea index (AHI) are the most important metrics in the diagnosis of sleep syndrome disease. In previous studies, these two tasks are usually implemented separately, which is both time- and resource-consuming. In this work, we propose a novel single electroencephalogram (EEG)-based collaborative learning network (EEG-CLNet) for simultaneous sleep staging and obstructive sleep apnea (OSA) event detection through multitask collaborative learning. The EEG-CLNet regards different tasks as a common unit to extract features from intragroups via both local parameter sharing and cross-task knowledge distillation (CTKD), rather than just sharing parameters or shortening the distance between different tasks. Our approach has been validated on two datasets with the same or better performance than other methods. The experimental results show that our method achieves a performance gain of 1%–5% compared with the baseline. Compared to previous works where two or even more models were required to perform sleep staging and OSA event detection, the EEG-CLNet could reduce the total number of model parameters and facilitate the model to mine the hidden relationships between different task semantic information. More importantly, it effectively alleviates the task bias problem in hard parameter sharing. As a consequence, this approach has notable potential to be a solution for a lightweight wearable sleep monitoring system in the future.

Bearing pitting, one of the common faults in mechanical systems, is a research hotspot in both academia and industry. Traditional fault diagnosis methods for bearings are based on manual experience with low diagnostic efficiency. This study proposes a novel bearing fault diagnosis method based on deep separable convolution and spatial dropout regularization. Deep separable convolution extracts features from the raw bearing vibration signals, during which a 3 × 1 convolutional kernel with a one-step size selects effective features by adjusting its weights. The similarity pruning process of the channel convolution and point convolution can reduce the number of parameters and calculation quantities by evaluating the size of the weights and removing the feature maps of smaller weights. The spatial dropout regularization method focuses on bearing signal fault features, improving the independence between the bearing signal features and enhancing the robustness of the model. A batch normalization algorithm is added to the convolutional layer for gradient explosion control and network stability improvement. To validate the effectiveness of the proposed method, we collect raw vibration signals from bearings in eight different health states. The experimental results show that the proposed method can effectively distinguish different pitting faults in the bearings with a better accuracy than that of other typical deep learning methods.

The Altay Prefecture in Xinjiang Uyghur Autonomous Region, China, boasts an abundance source of rock art, with the Dunde Bulaq rock art complex holding paramount significance in the study of local culture. This work employs hyperspectral techniques for data collection and analysis of the Dunde Bulaq I rock art site. Signal decomposition, pigment classification/clustering, and spectral curve analysis were employed to delve deeply into the information of the rock art hyperspectral data. Furthermore, cross-validation was utilized to reveal hidden information within rock art. To comprehend these hidden information, a more profound historical tracing was conducted, employing methods such as literature retrieval and image analysis, uncovering the internal relationship between the site I and the Chemurchek culture. This provides fresh material for studying the Bronze Age cultures within the Altai region.

Dual-inverter fed open-end winding permanent magnet synchronous motor (OW-PMSM) drive topology can extend the speed range under limited dc bus voltage, but it suffers from the zero-sequence current (ZSC) problem when with common bus configuration. Predictive current control (PCC) is a potential high-performance control method for electric drive. However, the PCC algorithm for dual-inverter system may bring a huge computational burden because of more space vectors (SVs) to traverse. In order to improve the steady-state current performance and suppress the ZSC, this article proposes a dual-vector PCC (DV-PCC) method for OW-PMSM with ZSC hysteresis control. The reference voltage vector is predicted in advance according the model. A novel sector judgment method is used to judge the sector and subsector where the reference voltage vector is located, and determines the two candidate SVs. Then the dwell time of the two candidate SVs is calculated through the projection length of the reference voltage vector. Moreover, the switching states redundancy feature of dual-inverter system is employed to suppress ZSC. As some SVs correspond to multiple switching states of different zero-sequence voltages (ZSVs), the switching state whose ZSV polarity is opposite to the ZSC polarity is selected as the output to realize the hysteresis control of ZSC. When the ZSVs of the candidate SVs contradict the hysteresis control demand, the zero-vector insertion method is used to guarantee the effective suppression of ZSC. The experiments are implemented in comparison with the existing DV-PCC method.

The combination of doped quantum dots (QDs) and graphene oxide (GO) results in QD–GO composites with enhanced light absorption properties. When exposed to light, the dual-emission Mn2+-doped ZnS QD–GO composite exhibits nonradiative energy transfer from photoexcited QDs to GO. Using steady-state spectroscopy and time-resolved spectroscopy, we find that the energy transfer efficiency between QDs’ surface defects and GO is higher than that between isolated Mn2+ions and GO. Moreover, the increase of GO concentration and reduction degree can improve the electron population of defect states of QDs, while the electron population in Mn2+ions is relatively less affected by GO. This study leads to a better understanding of the energy transfer dynamics between dual-emitting Mn2+-doped ZnS QDs and GO and should greatly contribute to the design of high-performance QD-based optoelectronic devices.

China is developing its BeiDou Satellite‐Based Augmentation System (BDSBAS) in accordance with international standards, and it has provided initial Single Frequency (SF) and Dual‐Frequency Multi‐Constellation (DFMC) services since July 2020. This paper gives a comprehensive introduction to BDSBAS from the aspects of its general design, system time datum, coordinate reference system and signal characteristics. The performance requirements and internationalization endeavors of BDSBAS also are presented. Additionally, the initial performance tests of BDSBAS SF service had been carried out based on the first two BDSBAS Geosynchronous Earth Orbit (GEO) satellites, including the quality of signal‐in‐space, and initial accuracy, integrity and availability. The results show that the characteristics of BDSBAS signal‐in‐space can well meet the requirements of the International Civil Aviation Organization (ICAO), and the initial performance of BDSBAS SF service can basically meet the requirements of Approach with Vertical guidance I (APV‐I). Finally, before the application of BDSBAS services into civil aviation uses, testing and certification will be organized by the Civil Aviation Administration of China (CAAC) and carried out in three steps. In the future, China will continue to promote the development of BDSBAS in a steady manner in order to provide aviation integrity services compatible with the ICAO standards in China and the surrounding areas as soon as possible.

The modification mechanism between nano-zinc oxide (nano-ZnO), polymerised styrene butadiene (PSB) and asphalt were studied by first-principles calculation, revealed the modification mechanism from the molecular scale. The asphalt was divided into four components: saturate, aromatic, resin and asphaltene, used nano-ZnO and PSB as modifiers for asphalt, constructed systems of asphalt modifiers and asphalt components, then geometric optimisation of each system. Density functional theory (DFT) was used to calculate the binding energy and the number of charges transferred between saturate, aromatic, resin and nano-ZnO, PSB. Calculated results show that nano-ZnO has chemical modification of hydroxylation to asphalt. Meanwhile, the nano-ZnO has a physical adsorption effect based on van der Waals force for each component of asphalt, which can be stably combined with asphalt. PSB can form aromatic-ring stacking with asphalt, making PSB combined with asphalt, the interaction of this accumulation is not strong, and the combination of PSB and asphalt is unstable. When nano-ZnO and PSB were simultaneously added into the asphalt, the absolute value of the binding energy between saturate, aromatic, resin and PSB all increased, nano-ZnO acts as a medium for attracting asphalt and PSB, the addition of nano-ZnO can effectively enhance the binding stability of PSB and asphalt.

This work focused on investigation of application properties of organoclay-modified natural rubber (NR) composites. This organoclay, flame-retardant (FR) dendrimer-modified organic montmorillonite (G-OMMT), was composed of dendrimer and montmorillonite. Different generations of G-OMMT were prepared. G3.0-OMMT was selected and FR composites were fabricated. Their tensile properties, thermal stability, flame-retardance, and other properties were compared. The tensile strength and elongation at break of NR/G3.0-OMMT-15 were 18.8 MPa and 586%, respectively, which were about 12.6% and 23.6% higher than that of pure NR, 16.7 MPa and 474%. In addition, when 10 phr of G3.0-OMMT was added, the wearing loss volume was 0.231 cm3, which was 42.4% lower compared with that of pure NR, 0.401 cm3. The NR composites also possessed better thermal stability. When the amount of this organoclay was 10 phr, T5%and T10%were 306.4°C and 349.5°C, respectively. They were 102.3°C and 29.2°C higher than that of pure NR. Moreover, the burning rate was decreased from 30.15 mm min−1for pure NR to 19.48 mm min−1for NR/G3.0-OMMT-20, and the limited oxygen index value was increased from about 18% for pure NR to about 28% for NR/G3.0-OMMT-20. Scanning electron microscopy analysis demonstrated that the layered silicates dispersed uniformly in these composites and possessed a good compatibility with the polymeric matrix.

Over the past decades, numerous experts have been investigated the correlation among distinct personality traits and job performance. However, relatively less attention was paid examining the significance of technological tools in sales management, especially in developing countries. This article explores the relationship among the five-factor model (FFM) of personality traits and sales performance (SP) with a moderating role of the technology. To this end, structural equation modeling and Fisher's Z transformation analysis were employed to analyze the hypotheses. The findings revealed that extraversion, conscientiousness, openness to experience, and emotional stability traits are positively correlated to SP. In contrast, agreeableness is not highly correlated with SP relatively than the remainder traits. Additionally, results revealed the significant effect of technology as a moderator which strengthens the association of FFM and SP. This study proposes diverse managerial implications and future directions for practitioners and academicians across the nations.