Search

• An unreported aramid fiber (AF)/PEEK composite is produced. • A simple surface modification method for the AF/PEEK composites is proposed. • A multiscale flexible-rigid CNT-COOH/ANF interphase has been constructed to improve interfacial adhesion. • The flexural strength (132.60 %) and ILSS (182.97 %) of AF/PEEK composites were enhanced. The application of aramid fiber (AF)/polyetheretherketone (PEEK) composites is currently hindered by the inert surface and poor wettability of AF, resulting in weak interfacial adhesion and poor mechanical properties. Surface coating and the introduction of nanostructures have been proven to be effective approaches to address this problem. Herein, a simple hybrid sizing agent has been developed to modify the AF surface, consisting of soluble polyimide (PI) as a compatibilizer, carboxyl-functionalized carbon nanotubes (CNT-COOH) as a rigid unit, and aramid nanofibers (ANF) as a flexible component. The synergetic effects of PI and the multiscale flexible-rigid structure (CNT-COOH/ANF) contribute to the formation of chemical and physical bonds between AF and PEEK matrix, further improving the interfacial adhesion and stress transfer efficiency. Attributed to the enhanced wettability and roughness of AF, compared with unsized AF, the flexural strength (220.97 MPa), modulus (13.26 GPa), ILSS (13.36 MPa), and storage modulus (12.93 GPa) of the AF/PEEK composite increase by 132.60 %, 99.00 %, 18.97 %, and 82.70 % respectively. Additionally, the flexible-rigid nanonetwork facilitates the penetration of the PEEK resin into pore spaces. This simple and effective approach exhibits promising potential in enhancing the interfacial bonding of AF/PEEK composites. [Display omitted] [ABSTRACT FROM AUTHOR]

AbstractTree-ring cellulose is a commonly used material for radiocarbon analysis. Extracting cellulose is labor-consuming and several devices that enable batchwise extraction have been developed. However, these devices bear the risk of sample contamination. The present study describes a new device which improves upon two aspects of currently available devices. First, to prevent cross-sample-contamination, we redesigned the drainage module to enable independent removal of chemical waste from each individual sample funnel. Second, we added covers to the sample funnels to reduce the risk of external contamination. Cellulose purity (i.e., holocellulose) was confirmed by Fourier Transform Infrared (FTIR) Spectroscopy. Furthermore, accuracy of the radiocarbon analysis was confirmed by results of 14C-blank samples and samples of known age. In conclusion, while maintaining labor-saving, our modified device significantly reduces the risk of sample contamination during extraction of tree-ring cellulose.

AbstractIn 2018, an Ionplus 200 kV MIni-CArbon DAting System (MICADAS) accelerator mass spectrometer (AMS) was installed at the Laboratory of AMS Dating and the Environment, Nanjing University (NJU-AMS Laboratory), China. The NJU-AMS Laboratory is largely devoted to research on radiocarbon dating and 14C analysis in fields of earth, environmental and archaeological sciences. The laboratory has successfully employed various pretreatment methods, including routine pretreatment of tree rings, buried wood and subfossil wood, seeds, charcoal, pollen concentrates, organic matter, and shells. In this study, operational status of the NJU-AMS is presented, and results of radiocarbon measurements made on different sample types are reported. Measurements on international standards, references of known age, and blank samples demonstrate that the NJU-AMS runs stably and has good reproducibility on measurement of single samples. The facility is capable of measuring 14C in samples with the precision and accuracy that meet the requirements for investigating annual 14C changes, history-prehistory age dating, and Late Quaternary stratigraphic chronology research.

Volatility of organic aerosols (OAs) significantly influences new particle formation and the occurrence of particulate air pollution. However, the relationship between the volatility of OA and the level of particulate air pollution (i.e., particulate matter concentration) is not well understood. In this study, we compared the chemical composition (identified by an ultrahigh-resolution Orbitrap mass spectrometer) and volatility (estimated based on a predeveloped parametrization method) of OAs in urban PM2.5(particulate matter with aerodynamic diameter = 2.5 µm) samples from seven German and Chinese cities, where the PM2.5concentration ranged from a light (14 µg m–3) to heavy (319 µg m–3) pollution level. A large fraction (71–98%) of compounds in PM2.5samples were attributable to intermediate-volatility organic compounds (IVOCs) and semivolatile organic compounds (SVOCs). The fraction of low-volatility organic compounds (LVOCs) and extremely low-volatility organic compounds (ELVOCs) decreased from clean (28%) to heavily polluted urban regions (2%), while that of IVOCs increased from 34 to 62%. We found that the average peak area-weighted volatility of organic compounds in different cities showed a logarithmic correlation with the average PM2.5concentration, indicating that the volatility of urban OAs increases with the increase of air pollution level. Our results provide new insights into the relationship between OA volatility and PM pollution levels and deepen the understanding of urban air pollutant evolution.

This paper studies when and under what conditions the actions undertaken by the power plants involved in China’s emission trading scheme (ETS) pilot became cost effective. Based on unique plant-level panel data and the difference-in-differences strategy, we identify that an insignificant initial reduction in cost efficiency occurred at the announcement stage for power plants in the pilot provinces; however, the cost efficiency of the pilot plants increased significantly following formal policy implementation. Additionally, the by-stage treatment effects differed across the pilot provinces due to localized market and non-market variations. Localized conditions of higher marketization, stricter policy enforcement, and lower carbon dependence enhanced this positive effect. The synthetic control results confirmed this variation in the policy effects. The carbon trading pilots resulted in improved efficiency in power plants in Shanghai, Guangdong, and Tianjin during the period 2013–2017, with an associated total cost saving of approximately 29.75 million RMB. To enhance the efficacy of the ETS policy, our findings suggest that the design of the policy should consider localized external factors.

Purpose: This study investigates the event-triggered vibration control problem for active full-vehicle suspension systems based on preview road information. The motivation is to design a novel event-triggered mechanism combined with preview information for improving vehicle stability while considering network-induced delays and limited communication bandwidth of networked control systems (NCSs). Methods: Firstly, as the payload of the vehicle is time varying, Takagi–Sugeno (T–S) algorithm is employed to characterize the vehicle suspension model. Secondly, a novel event-triggered function is designed to dynamically adjust the data transmission rate through preview information in the resource-limited communication network. Finally, since data transmission delays occur in the network system, an augmented Lyapunov–Krasovskii functional is employed to guarantee the suspension system stability with the H∞ standard. Results: The stability of the proposed event-triggered control system is proven. In addition, numerical and experimental details are presented to substantiate the feasibility of the event-triggered controller. Conclusion: The proposed event-triggered controller is more effectively to utilize the limited communication resource and improve the vehicle stability than the traditional control method.

Self-supervised learning (SSL) has gained significant traction within the remote sensing community, with pretraining a foundation model on large-scale unlabeled datasets for the interpretation of remote sensing images (RSIs) emerging as a trending direction. This approach aims to supplant the conventional practice of loading ImageNet pretrained weights, offering a more versatile and potentially more effective solution for handling RSIs. Among SSL techniques, contrastive learning excels in extracting general representations in the field of remote sensing. However, its excessive focus on inter-instance discrimination hinders the effectiveness of pretraining due to the diverse and complex geographical information present in RSIs. Moreover, the typical two-variations-as-one-pair pattern may be suboptimal, particularly given the temporal information specific to RSIs. In this article, we propose a novel method called promoting intra-instance similarity (PIS) for short, which leverages the temporal information specific to RSIs and increases the intra-instance variations to expand the positive representation space. Additionally, by PIS within this space, our foundation models develop the ability to extract more general and instance-invariant features that prove beneficial for various downstream tasks. Experiments show that our PIS method achieves state-of-the-art (SOTA) performance on ten datasets across four downstream remote sensing tasks, demonstrating the generalizability and efficacy of the proposed method. Through our preliminary investigation into intra-instance characteristics, we believe there exists substantial potential in this aspect, holding considerable promise for further exploration. The codes are available on the website: https://github.com/ShawnAn-WHU/PIS.git.

Hyperspectral image (HSI) classification is a core processing procedure in the remote sensing community, which has been recently well studied using vision transformers (ViTs). However, due to the high computational and memory complexities, existing transformer-based classification methods tend to restrict the spatial extent of the transformer to small cropped HSI patches instead of the whole HSI data, thus sacrificing the essential strength of transformers in long-range interaction modeling and overlooking the beneficial multiscale features in HSI data. Inspiringly, here we propose PSFormer, a novel pyramid superpixel transformer (PSFormer) method specifically for HSI classification, in order to make full use of the transformer to excavate multiscale local-global features in HSI data. Specifically, a progressive superpixel merging strategy is introduced to flexibly control the scale of feature maps. Furthermore, a unique transformer backbone design based on a spectral attention layer and a classification head with a gate mechanism are developed, to adaptively exploit valuable local-global information at different scales with low computational cost. Extensive experimental results on five widely used datasets demonstrate the superiority of PSFormer over other state-of-the-art networks. For the sake of reproducibility, the related code of the PSFormer method will be open-sourced at: https://github.com/immortal13.

In addressing the insufficiencies of feature insertion, inaccurate positioning, and incompatible feature sizes in data augmentation algorithms based on deep learning for detecting microscopic defects on printed circuit boards (PCBs), this paper proposes a novel approach incorporating multiple strategies for small target alignment insertion. First, traditional linear feature extraction methods are transformed into a multiscale comprehensive analysis process. Subsequently, point-to-point matching calculations are converted into region-wise traversals to enhance accuracy and constrain the matching region. Next, geometric correspondences are determined through the computation of a transfer matrix, effectively eliminating perspective distortions. Finally, by constructing a top-down pyramid optical flow module, size limitations are overcome while enhancing features of small target defects. Experimental results demonstrate that this method significantly improves the recognition accuracy of the network model for small target defects on PCB surfaces.

Addressing the challenges of complex backgrounds, minute defects, and irregular shapes in PCB defect images that often lead to missed detections, inaccurate localizations, and false positives, this article introduces an improved you only look once (YOLO) model, termed YOLO-Biformer, to enhance the network’s ability to detect surface defects on PCBs. First, YOLO-Biformer incorporates a hybrid attention module to differentiate the importance among various channels, thus strengthening the extraction of small target defect features and preventing the loss of minor target information often caused by deep convolutional networks. Second, the model introduces a jumping hollow space convolutional pyramid aimed at preserving more image details and interrelated information, thereby boosting the network’s defect localization capability. Finally, the Enhanced Intersection Over Union-FocalLoss loss function is employed to enhance the network’s ability to distinguish between defects with similar features. Experimental results demonstrate that the proposed algorithm increases the mean average precision by 4.1%, showing excellent performance in recognizing small target defects on printed circuit board surfaces and achieving both high accuracy and real-time capabilities.

Deep learning-based building instance extraction on remote sensing imagery (RSI) has achieved tremendous success under the large-scale labeled training data. However, multitarget domain adaptation building instance extraction (MD-BIE) is still a challenge task that involves transferring knowledge from a source domain to multiple unlabeled target domains, which poses various semantic gaps between and within multiple domains, e.g., style, illumination, resolution, density, and scale. Most current methods for single-target domain adaptation are not applicable to the more realistic MD-BIE task. To this end, we propose a novel domain-common approximation learning (DAL) for both modeling intradomain and interdomain adaptation, thus obtaining robust MD-BIE. DAL contains three main modules: multidomain style transfer (MST), multidomain feature approximation (MFA), and multidomain cascaded instance extraction (MCIE). To alleviate the semantic gaps between multiple domains for interdomain adaptation, we first employ the MST to learn multiple target-domain-like features that preserve both the styles of target domains and the content of the source domain, and then use the MFA to approximate these features toward a central domain-common space, thus producing domain-common semantic representations. Moreover, we develop the MCIE with hierarchical extraction losses for intradomain adaptation to extract precise building instance contours from the domain-common semantic representations, further eliminating the potential gaps within multiple domains. By colearning these three modules in an end-to-end manner, the DAL bridges the semantic gaps between and within multiple domains. Extensive experiments on different popular MD-BIS tasks (SAB $\rightarrow $ Crowd & WHU, Crowd $\rightarrow $ SAB & WHU, SAB $\rightarrow $ Crowd & SAB & WHU, and SAB $\rightarrow $ WHU) show that our DAL outperforms the current methods by a significant margin.

Subspace clustering has achieved remarkable performance for hyperspectral image (HSI). However, existing methods are often computationally expensive and have limited ability to capture the intrinsic structural information of HSI. In this article, we propose a structural prior-guided subspace clustering method, which simultaneously incorporates the local and nonlocal spatial information and the cluster prior information. Accordingly, three efficient regularizations are developed. Considering the local connectivity of pixels, we propose an $\ell _{2,1}$ norm-based constraint on the representation difference matrix to improve the homogeneity of clustering result. Next, to capture the nonlocal geometric structure of HSI, we propose a manifold-based regularization with an adaptively learned landmark graph. Furthermore, we explore the block-diagonal cluster structure of HSI and develop a landmark-based clustering constraint, which makes the representations more favorable for clustering. Our local constraint is imposed on all the data points due to its efficiency and the latter two are solely imposed on landmarks, leading to computationally efficient regularizations. Due to the local constraint, the manifold and cluster structure of the landmarks can be effectively propagated to all the data points. To make our model scalable to large-scale data, we learn a compact dictionary with an orthogonal constraint, significantly reducing the number of parameters. In addition, we propose a novel landmark selection method to support our landmark-based constraints using multiscale super-pixel segmentation and clustering, which improves the uniformity and diversity of landmarks. We also develop an efficient algorithm to solve the proposed model. Experimental results demonstrate that our model outperforms the state-of-the-art.

Image-level weakly supervised semantic segmentation (WSSS) methods have gained prominence in remote sensing image building extraction tasks, primarily due to their cost-effectiveness in manual annotation. However, owing to the intricate details present in remote sensing building images, the pseudolabels generated from existing image-level weakly supervised methods often encounter issues of incorrect activation and unclear boundaries. In this article, we propose a global-to-local dynamic information enhancement (G2LDIE) framework. This framework effectively extracts global information from remote sensing building images and supplements local details through a local information enhancement (LIE) module, generating more accurate pseudolabels. Additionally, we propose a dynamic label guide strategy (DLGS) to enhance model consistency in category representation across various scale images. To address the challenge of unclear building boundaries issue in pseudolabels, we introduce a segment anything model (SAM) postprocessing (SPP) method, which can better correct the boundaries of building images at different resolutions while reducing computational costs. Extensive and detailed experiments on three datasets confirm that our framework can generate refined pseudolabels and outperform other image-level weakly supervised methods in terms of accuracy and generalization performance in building extraction.

Haze pollution in hyperspectral images (HSIs) leads to surface information lack and image clarity degradation, which seriously affects the performance of subsequent image interpretation. Existing model-based hyperspectral haze removal methods enjoy good interpretability and generalization, but they can only process images in a specific wavelength range due to the principle limitation. Deep learning-based dehazing methods have good feature extraction capability, but the cost of acquiring sufficient training data is high in practical applications. At the same time, taking into account that HSIs have spectral low-rank structures, fully utilizing the low-rank property will facilitate the reconstruction of HSIs. In order to combine the complementary benefits of deep learning-based and physical model-based approaches, we decide to formulate HSI dehazing reconstruction as an unsupervised deep image prior (DIP) framework. Specifically, we propose an unsupervised dehazing network with hybrid prior constraints (HPC-UDN) for HSI haze removal, which effectively integrates low-rank prior, deep priors, and physical haze prior. First, the low-rank prior of hyperspectral data is characterized by matrix decomposition, where the decomposition factors are learned through two generative networks. Then, multiple spectral groups are divided based on the correlation and complementarity between spectral bands. In order to exchange information between adjacent spectral groups, a novel spectral grouping feature fusion (SGFF) module is designed, which connects neighboring spectral groups to transfer spectral and spatial features. Finally, high-quality HSI is recovered by merging the features extracted from each spectral group. Extensive simulated and real-data experiments certify the effectiveness and robustness of the presented unsupervised approach and potential applications in the GF-5 image dehazing task.

Hyperspectral image (HSI) qualities are limited by a mixture of Gaussian noise, impulse noise, stripes, and deadlines during the sensor imaging process, resulting in weak application performance. To enhance HSI qualities, methods based on convolutional neural networks have been successively applied to restore clean data from the observed data. However, the architecture of these methods lacks spectral and spatial constraints, and the convolution operators have limited receptive fields and inflexible model inferences. Thus, in this study, we propose an efficient end-to-end transformer, named HSI denoising transformer (Hider), for mixed HSI noise removal. First, a U-shaped 3-D transformer architecture is built for multiscale feature aggregation. Second, a multihead global spectral attention module within the spectral transformer block is designed to excavate information in different spectral patterns. Finally, an additional locally enhanced cross-spatial attention module within the spatial–spectral transformer block is constructed to build the long-range spatial relationship to avoid the high computational complexity of global spatial self-attention. Through the imposition of global correlations along spectrum and spatial self-similarity constraints on the transformer, our proposed Hider aims to capture long-range spatial contextual information and cluster objects with the same spectral pattern for HSI denoising. To verify the effectiveness and efficiency of Hider, we conducted extensive simulated and real experiments. The denoising results on both simulated and real-world datasets show that Hider achieves superior evaluation metrics and visual assessments compared with other state-of-the-art methods.

Tsampa may contain pesticide residues and mycotoxins, which may pose a risk to human health. Currently, pesticide detection and mycotoxin detection are two independent experiments. To improve the efficiency of the analysis, a method based on QuEChERS combined with ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) for the simultaneous determination of 78 pesticides and 16 mycotoxins in tsampa was developed. All the target compounds showed good linear correlation with correlation coefficients (R2) greater than 0.9990. The limits of detection (LODs) and limits of quantification (LOQs) were in the ranges of 0.10–3.00 μg kg−1and 0.40–10.00 μg kg−1, respectively. The average recoveries of the pesticides and mycotoxins spiked at the 1, 2, and 10-fold LOQ were in the range of 73.0–115.2%, and the relative standard deviations (RSDs) were lower than 11.7%. This method was applied to 19 batches of real samples in which 32% of samples exceeded the maximum residue limits of the European Union involving aflatoxin G2, ochratoxin A, and hexaconazole. It proved to be excellent, efficient, greatly simplified, and highly applicable, which could reduce the workload and time significantly for the daily monitoring of the pesticides and mycotoxins in tsampa.

Landsat optical sensor is crucial for long-term observations of the Earth’s surface with a 30-m spatial resolution. However, the 16-day revisit cycle and severe atmospheric interference have impeded the monitoring of rapid surface changes. Spatiotemporal fusion (STF) is a classic method of predicting Landsat surface reflectance with multitemporal and multisource data, but it is limited by unpredictable temporal changes and cloudy Landsat–moderate resolution imaging spectroradiometer (MODIS) image pairs. Another emerging solution is synthetic aperture radar (SAR)-to-optical image translation (S2OIT), which always produces spectral distortions. To tackle these defects, we propose a new data-driven solution, SAR–optical data-based spatial–spectral fusion (SOSSF), which combines the high-spatial and cloud-free advantages of Sentinel-1 data and the high-spectral and high-temporal advantages of MODIS images to synthesize high-spatial and high-temporal Landsat-8 images. To achieve this solution, we first establish a worldwide benchmark dataset, namely, Sentinel-1, MODIS, and Landsat-8 triplets (SMILE), with various land cover types and all meteorological seasons, satisfying the big data requirements of deep learning (DL). Second, we design an attention-based dual-path fusion network (ADFNet) to respectively extract and fully fuse spatial and spectral information from SAR–optical data. Extensive experiments suggest that the proposed SOSSF solution outperforms the state-of-the-art (SOTA) STF and S2OIT solutions, robustly performing in the continuously changing and frequently cloudy regions. The proposed ADFNet model achieves the best visual effect and the highest accuracy in different scenes, seasons, and bands. Furthermore, the proposed SOSSF solution is proven to be a practical way to simulate time-series and large-scale Landsat-8 surface reflectance, considerably enriching raw Landsat-8 products.

Building extraction aims to segment building pixels from remote sensing (RS) images and plays an essential role in many applications, such as city planning and urban dynamic monitoring. Over the past few years, deep learning methods with encoder–decoder architectures have achieved remarkable performance due to their powerful feature representation capability. Nevertheless, due to the varying scales and styles of buildings, conventional deep learning models always suffer from uncertain predictions and cannot accurately distinguish the complete footprints of the building from the complex distribution of ground objects, leading to a large degree of omission and commission. In this article, we realize the importance of uncertain prediction and propose a novel and straightforward uncertainty-aware network (UANet) to alleviate this problem. Specifically, we first apply a general encoder–decoder network to obtain a building extraction map with relatively high uncertainty. Second, in order to aggregate the useful information in the highest-level features, we design a prior information guide module (PIGM) to guide the highest-level features in learning the prior information from the conventional extraction map. Third, based on the uncertain extraction map, we introduce an uncertainty rank algorithm (URA) to measure the uncertainty level of each pixel belonging to the foreground and the background. We further combine this algorithm with the proposed uncertainty-aware fusion module (UAFM) to facilitate level-by-level feature refinement and obtain the final refined extraction map with low uncertainty. To verify the performance of our proposed UANet, we conduct extensive experiments on three public building datasets, including the WHU building dataset, the Massachusetts building dataset, and the Inria aerial image dataset. Results demonstrate that the proposed UANet outperforms other state-of-the-art (SOTA) algorithms by a large margin. The source code of the proposed UANet is available at https://github.com/Henryjiepanli/Uncertainty-aware-Network.

Acetic acid (AC) is a major by-product from fermentation processes for producing succinic acid (SA) using Actinobacillus succinogenes. Previous experiments have demonstrated that sodium bisulfate (NaHSO3) can significantly decrease AC production by A. succinogenesGXAS137 during SA fermentation. However, the mechanism of AC reduction is poorly understood. In this study, the transcriptional profiles of the strain were compared through Illumina RNA-seq to identify differentially expressed genes (DEGs). A total of 210 DEGs were identified by expression analysis: 83 and 127 genes up-regulated and down-regulated, respectively, in response to NaHSO3treatment. The functional annotation analysis of DEGs showed that the genes were mainly involved in carbohydrates, inorganic ions, amino acid transport, metabolism, and energy production and conversion. The mechanisms of AC reduction might be related to two aspects: (i) the lipoic acid synthesis pathway (LipA, LipB) was significantly down-regulated, which blocked the pathway catalyzed by pyruvate dehydrogenase complex to synthesize acetyl-coenzyme A (acetyl-CoA) from pyruvate; (ii) the expression level of the gene encoding bifunctional acetaldehyde-alcohol dehydrogenase was significantly up-regulated, and this effect facilitated the synthesis of ethanol from acetyl-CoA. However, the reaction of NaHSO3with the intermediate metabolite acetaldehyde blocked the production of ethanol and consumed acetyl-CoA, thereby decreasing AC production. Thus, our study provides new insights into the molecular mechanism of AC decreased underlying the treatment of NaHSO3and will deepen the understanding of the complex regulatory mechanisms of A. succinogenes.

The flow characteristics in reciprocating pump-pipe system are inevitably subject to the interaction effects between pump and pipeline. In this paper, a parametric interaction model of flow characteristics in reciprocating pump-pipe system is presented by considering the coupling effect and leakage characteristics. The analytical solutions of the proposed interaction equations are deduced with the aid of the harmonic balance method. A testing platform of a simplex plunger reciprocating pump-pipe system is employed for validating the proposed model. The experimental tests under different operating conditions are used to compare the pressure pulsations and discharge flow rates from analytical solutions. It is observed that the analytical solutions are in good agreement with the test data and numerical results, and the prediction error is within 13%, thus validating the presented interaction model and analytical solutions. The interaction effects as well as the flow characteristics in reciprocating pump-pipe system are analyzed and discussed using the proposed model. It is found that the opening of flow control valve has remarkable influence on the relationship between flow rate and operating speed of reciprocating pump. The pressure pulsation within a pipe is closely related to and significantly affected by the plunger-to-cylinder clearance and the plunger stroke.

Band selection (BS) reduces effectively the spectral dimension of a hyperspectral image (HSI) by selecting relatively few representative bands, which allows efficient processing in subsequent tasks. Existing unsupervised BS methods based on subspace clustering are built on matrix-based models, where each band is reshaped as a vector. They encode the correlation of data only in the spectral mode (dimension) and neglect strong correlations between different modes, i.e., spatial modes and spectral mode. Another issue is that the subspace representation of bands is performed in the raw data space, where the dimension is often excessively high, resulting in a less efficient and less robust performance. To address these issues, in this article, we propose a tensor-based subspace clustering model for hyperspectral BS. Our model is developed on the well-known Tucker decomposition. The three factor matrices and a core tensor in our model encode jointly the multimode correlations of HSI, avoiding effectively to destroy the tensor structure and information loss. In addition, we propose well-motivated heterogeneous regularizations (HRs) on the factor matrices by taking into account the important local and global properties of HSI along three dimensions, which facilitates the learning of the intrinsic cluster structure of bands in the low-dimensional subspaces. Instead of learning the correlations of bands in the original domain, a common way for the matrix-based models, our model learns naturally the band correlations in a low-dimensional latent feature space, which is derived by the projections of two factor matrices associated with spatial dimensions, leading to a computationally efficient model. More importantly, the latent feature space is learned in a unified framework. We also develop an efficient algorithm to solve the resulting model. Experimental results on benchmark datasets demonstrate that our model yields improved performance compared to the state-of-the-art.

Semantic segmentation is one of the basic tasks in computer vision. Its purpose is to achieve pixel-level scene segmentation. With the popularity of depth sensors, combining depth data with RGB images for semantic segmentation can improve the accuracy of semantic segmentation. First, this paper mainly summarizes the fusion of RGB information and depth information and then describes the RGBD semantic segmentation method, evaluation metrics, data set, and comparison of the results on the two mainstream data sets, and then make a prospect of possible future research directions, and finally, a conclusion is made. This part of the work has a certain guiding significance for future research on RGBD semantic segmentation and lays a foundation for later research.

Conversion of waste wood containing urea–formaldehyde resin (WW-UF) into solid fuel with lower nitrogen content and better fuel property through the hydrothermal carbonization (HTC) process was studied. With the increase of HTC temperature from 180 to 260 °C, the mass yield and energy yield of the hydrochar decreased from 81.2 to 38.7% and from 81.6 to 62.3%, respectively, while the higher heating values (HHV) of the hydrochar increased from 19.2 to 27.1 MJ/kg. Due to the hydrolysis of urea–formaldehyde resin into urea, the denitrogenation efficiency could reach over 80% at 180 °C. However, the denitrogenation efficiency decreased at elevated HTC temperature because of the condensation/polymerization of urea with carbohydrate derivatives. The combustion and pelletization behaviors of the hydrochar were all improved compared to that of the raw WW-UF. FT-IR spectra confirmed that the UF resin and hemicellulose in the WW-UF could be hydrolyzed at 180 °C, but carbonization of cellulose occurred with HTC temperature over 240 °C. The process water was analyzed by LC–MS/MS to confirm that the mono-saccharides, carboxylic acids, carbocyclic compounds, O-heterocyclic compounds, and N-heterocyclic compounds were presented in process water. The study confirmed that hydrothermal carbonization could be one promising method to covert the WW-UF into clean and efficient solid fuel.

Plant oil adjuvants are widely used to improve the utilization rate of pesticides. In this study, the uptake, translocation, and terminal residue of chlorantraniliprole and difenoconazole spraying with plant oil adjuvant in rice (Oryza sativa L.) were evaluated. After being mixed with the tank-mixed plant oil adjuvant, the cuticular wax of rice leaf was destroyed, which decreased the hydrophobicity of the rice leaf and facilitated the wetting, spreading, and penetration of pesticides onto the rice leaf. Additionally, the adjuvant promoted the translocation of difenoconazole from leaves to stems, but had little effect on the translocation of difenoconazole from leaves to roots, while inhibiting chlorantraniliprole translocation. Although adjuvant increased the initial deposition of chlorantraniliprole and difenoconazole on rice, the terminal residue was not significantly affected. The findings can promote the safe use of chlorantraniliprole and difenoconazole in rice production, especially when used with plant oil adjuvants. In the future, studies on more rice cultivars will be necessary to determine the generality of the conclusions.

ABSTRACTRab-interacting lysosomal protein – like 2 (RILPL2) has been reported to be associated with prognosis and tumor biological functions in breast cancer and endometrial carcinoma. However, its expression and functional role in non-small cell lung cancer (NSCLC) remain unclear. The expression and clinical data of lung adenocarcinoma (LUAD) and lung squamous carcinoma (LUSC) were downloaded from the TCGA database. The expression of RILPL2 in NSCLC cell lines was verified by the Western blot. We used online databases and bioinformatics analysis tools to explore its prognostic value, potential biological functions, and correlations with tumor immune microenvironment.The expression of RILPL2 was significantly lower in NSCLC compared with adjacent normal tissues. Low RILPL2 expression was associated with worse overall survival (OS) in NSCLC. The GO analysis showed RILPL2 was comprehensively involved in immune activity. RILPL2 expression was significantly positively correlated with the infiltration levels of B cells, CD8+T cells, CD4+T cells, macrophages, neutrophils, dendritic cells (P < 0.001), and it was also significantly positively correlated with programmed cell death ligand 1 (PD-L1/CD274) (P < 0.001). High RILPL2 expression could predict better immunotherapy response and prognosis in the immunotherapy cohort. The GSEA analysis showed low RILPL2 expression was associated with glycolysis process in LUAD, which was verified in vitro.These results showed RILPL2 expression was correlated with prognosis, tumor microenvironment, and immunotherapy response in NSCLC. Besides, RILPL2 may regulate glycolysis in LUAD.

Background:: Diabetic wound healing remains a major challenge due to the impaired functionality of angiogenesis by persistent hyperglycemia. Mesenchymal stem cell exosomes are appropriate candidates for regulating the formation of angiogenesis in tissue repair and regeneration. Here, we explored the effects of exosomes derived from human amniotic mesenchymal stem cell (hAMSC-Exos) on the biological activities of human umbilical vein endothelial cells (HUVECs) treated with high glucose and on diabetic wound healing and investigate lncRNAs related to angiogenesis in hAMSC-Exos. Methods:: hAMSCs and hAMSC-Exos were isolated and identified by flow cytometry or western blot. A series of functional assays such as cell counting kit-8, scratching, transwell and tube formation assays were performed to evaluate the potential effect of hAMSC-Exos on high glucose-treated HUVECs. The effect of hAMSC-Exos on diabetic wound healing were tested by measuring wound closure rates and immunohistochemical staining of CD31. Subsequently, the lncRNAs profiles in hAMSC-Exos and hAMSCs were examined to screen the lncRNAs related to angiogenesis. Results:: The isolated hAMSC-Exos had a size range of 30–150 nm and were positive for CD9, CD63 and CD81. The hAMSC-Exos facilitate the functional properties of high glucose-treated HUVECs including the proliferation, migration and the angiogenic activities as well as wound closure and angiogenesis in diabetic wound. hAMSC-Exos were enriched lncRNAs that related to angiogenesis, including PANTR1, H19, OIP5-AS1 and NR2F1-AS1. Conclusion:: Our findings demonstrated hAMSC-Exos facilitate diabetic wound healing by angiogenesis and contain several exosomal lncRNAs related to angiogenesis, which may represent a promising strategy for diabetic wound healing.

In recent years, solar-powered photocatalytic water splitting for hydrogen production has drawn great attention as a method to supply clean energy and reduce environmental pollution. This approach, driven by solar energy to synthesize hydrogen gas, is completely renewable and can be used to power hydrogen fuel cells to effectively solve the global energy problem, while only releasing environmentally benign water. Accordingly, popularizing this subject and educating students with the knowledge of this powerful technology are urgently needed. In this paper, Pt-loaded commercial titanium dioxide (P25) is used as the photocatalyst to explore the process of photocatalytic hydrogen evolution. A simple experimental demonstration is presented, integrating photocatalytic hydrogen production with a hydrogen fuel cell system in chemistry class. This experiment is suitable for use with undergraduate students as a systematic and basic laboratory experience or as a demonstration for primary/middle/high school students to stimulate their research interests in field of renewable energy. The principles and reaction mechanisms involved in both photocatalytic water splitting and hydrogen fuel cell operation, as well as the experimental design to link photocatalysis with fuel cells, are systematically discussed.

The safety of traditional Chinese medicine (TCM) has garnered considerable interest worldwide. In this study, a high-throughput method for the determination of 255 pesticide residues in decoctions of Radix Codonopsis and Angelica sinensiswas developed using liquid chromatography-time-of-flight/mass spectrometry. The methodological verification demonstrated the accuracy and reliability of this method. The frequently detected pesticides in Radix Codonopsis and Angelica sinensiswere determined to build a correlation between pesticide properties and the transfer rate of pesticide residues in their decoctions. Water solubility (WS) with a higher correlation coefficient (R) made a significant contribution to the accuracy of the transfer rate prediction model. The regression equations for Radix Codonopsis and Angelica sinensiswere T= 13.64 log WS + 10.56 with a correlation coefficient (R) of 0.8617 and T= 10.66 log WS + 25.48 with a correlation coefficient (R) of 0.8072, respectively. This study provides preliminary data on the potential risk of exposure to pesticide residues in Radix Codonopsis and Angelica sinensisdecoctions. Furthermore, as a case study on root TCM, this approach may serve as a model for other TCMs.

Change detection (CD), as one of the central problems in Earth observation, has attracted a lot of research interest over recent decades. Due to the rapid development of satellite sensors in recent years, we have witnessed an enrichment of the CD source data with the availability of very-high-resolution (VHR) multispectral imagery, which provides abundant change clues. However, precisely locating real changed areas still remains a challenge. In this article, we propose an end-to-end superpixel-enhanced CD network (ESCNet) for VHR images, which combines differentiable superpixel segmentation and a deep convolutional neural network (DCNN). Two weight-sharing superpixel sampling networks (SSNs) are tailored for the feature extraction and superpixel segmentation of bitemporal image pairs. A UNet-based Siamese neural network is then employed to mine the different information. The superpixels are then leveraged to reduce the latent noise in the pixel-level feature maps while preserving the edges, where a novel superpixelation module is used to serve this purpose. Furthermore, to compensate for the dependence on the number of superpixels, we propose an innovative adaptive superpixel merging (ASM) module, which has a concise form and is fully differentiable. A pixel-level refinement module making use of the multilevel decoded features is also appended to the end of the framework. Experiments on two public datasets confirmed the superiority of ESCNet compared to the traditional and state-of-the-art (SOTA) deep learning-based CD (DLCD) methods.

Feature matching is a fundamental and critical task for remote sensing image registration (RSIR). However, numerous outliers (false matches) harm the feature point neighborhood structure due to the view transformation from the camera. Meanwhile, unknown local distortion obscures the distinction between inliers (correct matches) and outliers. To solve these problems, we propose a global–local consistency network (GLoCNet) for feature matching to exclude the interference of outliers under various transformation patterns and provide stable neighborhood support for the similarity metric of feature points. Specifically, a global transformation consistency (GTC) module is proposed to obtain a neighbor pool by exploiting the compact nature of the inlier distribution under different transformation patterns. In addition, feature points interact with information through the local neighborhood consistency (LNC) module in center-based graph construction. Finally, the difference between inliers and outliers is increased by dynamically adjusting the upper limitation distance of the outlier and suppressing its effect in the neighborhood. We conducted rich experiments on extensive datasets to verify the effectiveness of the proposed method. The experimental results illustrate that the proposed GLoCNet can effectively handle numerous outliers and achieve satisfactory registration results under various transformation patterns.

Hyperspectral image (HSI) classification with limited training samples has been well studied in recent years. Among them, the few-shot learning (FSL) technique demonstrates excellent processing capability under limited labeled samples. Nevertheless, the current FSL-based works provide scarce attention to effective class prototypes and metric types, resulting in high generalization error and poor interpretation during the cross-domain testing phase. A dual-adjustment mode-based cross-domain meta-learning (DMCM) method for few-shot HSI classification is proposed to tackle this issue. Specifically, a three-dimensional ghost attention network (TGAN) with strong learning capability without massive parameters is first constructed. Meanwhile, a dual-adjustment mode comprising intracorrection (IC) and interalignment (IA) learning strategies is then adopted to solve domain shift issue via episode-level meta-tasks, where IC and IA focus on effective class prototypes and data distribution differences between domains, respectively. Afterward, considering that the traditional Euclidean distance metric is insensitive to the distribution of within-class samples, the class-covariance metric (CCM) is employed to account for the distribution in feature space of each class to optimize decision boundary and alleviate the misclassification problem. Extensive experiments on three publicly available target hyperspectral datasets demonstrate the effectiveness of the proposed method in comparison with other state-of-the-art (SOTA) methods. The codes will be available on the website at https://github.com/HlEvag/DMCM.git.

AbstractIn the present study, conditions for succinic acid (SA) production using xylose mother liquor (XML) as culture medium by Actinobacillus succinogenesGXAS137 were optimized. Firstly, single–factor experiments were performed to evaluate the basal culture medium for SA fermentation. Thereafter, the Plackett–Burman design was used to screen out three significant factors of XML, corn steep liquor powder (CSLP) and MgCO3affecting the SA yields from the original nine factors. Subsequent use of steepest ascent experiment determined the center area of the three factors. Finally, the response surface methodology was used to further optimize the interactions between the three main factors and predict the maximum SA concentration through Box-Behnken design. The optimal conditions of SA fermentation were maximally documented in the XML (110 g/L), CSLP (18.86 g/L) and MgCO3(69.12 g/L). The maximum production of SA was 58.06 ± 0.57 g/L after 60 h with a yield of 0.72 ± 0.06 g/g total sugar, approaching the predicted value (57.99 g/L). It was 1.63-fold of the SA production obtained with the basic medium (35.54 g/L). In addition, batch fermentations were carried out in a 1.3-L stirred bioreactor and SA reached 58.47 g/L. These results indicate that XML could be an alternative substrate for the economical production of SA by A. succinogenesGXAS137.

Tone is the changing trend of pitch with time. In Chinese, tone plays an essential role for distinguishing meaning. Chinese dialect’s tone is more complex with Mandarin. In the field of Chinese dialect phonetics research, using human earing to recognize the types of tones is still the main method. So batch processing is not possible. In this paper, we construct a GSNP (gated spiking neural P) model with 2 layers which can process time series data to recognize the tones of Chinese dialects. The average accuracy rate of seven cities’ speech is more than 97%. Even in the case of small training samples, compared with other methods, the GSNP model has simpler structure, higher accuracy and more efficiency. It can not only improve the work efficiency of Chinese dialect field investigation, but also help researchers to screen the sounds with special sounds.