Search

The demand for underwater pressure sensitive adhesives (PSAs) is rapidly increasing in fields such as underwater engineering and biomedicine. However, the achievement of underwater adhesion of PSAs remains a challenge because of the hydration layer that hinders the interaction between the adhesive and the substrate. Herein, a new type of underwater PSA was synthesized by the copolymerization of hydrophobic unsaturated poly(1,2-butylene oxide) (UPBO) and hydrophilic itaconic acid monomers using solvent-free ultraviolet curing. The PSA has demonstrated substrate-independent underwater adhesion strengths ranging from 108 to 141 kPa on both hydrophilic (glass, wood, steel) and hydrophobic (PET, PMMA, PTFE) substrates. The underwater adhesion performance of PSA remains stable during 30 adhesion–detachment cycles and incubation in water for 20 days. Notably, PSA shows cytocompatibility, antimicrobial, and degradable properties and can be used for rapid hemostasis of skin wounds. Experimental characterizations confirm that the process of underwater adhesion is achieved by hydrophobic alkyl side chains of the PBO chain segments, which repel water at the adhesive–substrate interface. This study should provide both practical and facile design strategies for multifunctional underwater PSAs that can be used in a variety of applications.

The human calcium-sensing receptor (CaSR) detects fluctuations in the extracellular Ca2+concentration and maintains Ca2+homeostasis1,2. It also mediates diverse cellular processes not associated with Ca2+balance3–5. The functional pleiotropy of CaSR arises in part from its ability to signal through several G-protein subtypes6. We determined structures of CaSR in complex with G proteins from three different subfamilies: Gq, Giand Gs. We found that the homodimeric CaSR of each complex couples to a single G protein through a common mode. This involves the C-terminal helix of each Gα subunit binding to a shallow pocket that is formed in one CaSR subunit by all three intracellular loops (ICL1–ICL3), an extended transmembrane helix 3 and an ordered C-terminal region. G-protein binding expands the transmembrane dimer interface, which is further stabilized by phospholipid. The restraint imposed by the receptor dimer, in combination with ICL2, enables G-protein activation by facilitating conformational transition of Gα. We identified a single Gα residue that determines Gqand Gsversus Giselectivity. The length and flexibility of ICL2 allows CaSR to bind all three Gα subtypes, thereby conferring capacity for promiscuous G-protein coupling.

Magnesium (Mg) is abundant in humans. Studies have shown that Mg2+significantly affects physiological processes such as osteogenesis, osteoblast adhesion and motility, immunomodulation, and angiogenesis. Mg-based alloy porous scaffolds have attracted increasing attention because of their degradability and mechanical properties. Hence, wire-cut EDM was used to fabricate porous scaffolds. In addition, fluoride treatment afforded an MgF2coating, and racemic polylactic acid PDLLA was used as the carrier of rhBMP-2, which was evenly coated on the surface of the passivated porous alloy scaffold to slowly release growth factors and slow down the degradation rate. The rat femoral condyle defect model experiment was performed to study the in vivobone regeneration capacity of porous scaffolds and compare the differences in the healing effect with or without rhBMP-2. The enhanced corrosion resistance of the porous scaffold was confirmed through in vitroimmersion experiments. Micro-CT implied that porous scaffold with rhBMP-2 induced new bone formation and the new bone formation along the pores, as well as the histological examination in vivo. In summary, the porous scaffold promotes bone formation and has great potential for clinical translation.

AbstractIn this study, a series of polyethylene glycol (PEG)/hydrochloric acid-modified sepiolite (HSEP) composite phase change materials (PCMs) are fabricated viavacuum impregnation. HSEP exhibits high adsorption capacity, rendering it superior to natural sepiolite as carriers for PEG. The resulting composite PCMs possess a melting enthalpy of up to 88.9 J/g and maintain stable thermal performances and chemical structures over 100 heating–cooling cycles between room temperature and 65 °C, thus, indicating long-term reliability. Calorimetry studies on cement paste containing 30% composite PCMs reveal a 24.14% reduction in 3-day cumulative hydration heat. However, the mechanical strength and thermal conductivity of the cement paste are adversely affected. Hence, carbon fibers (CFs) are introduced as reinforcement, resulting in a 28-day compressive strength of 45.6 MPa for cement paste containing 20% composite PCMs and 0.6% CFs. The fabricated composite PCMs are promising functional materials for hydration heat control and energy storage in concrete structures.

Convolutional neural networks (CNNs) have been widely used in hyperspectral image (HSI) classification due to their ability to extract image features effectively. However, under the condition of limited samples, the modeling ability of CNNs for the relationships among samples is limited. At present, research on the classification of HSIs with a small number of samples remains an important challenge in the field of HSI processing. Recently, graph convolutional networks (GCNs) have been applied in HSI classification tasks. In this article, a multihop graph rectifies attention and spectral overlap grouping convolutional fusion network (MRSGFN) for HSI classification is proposed. In the graph convolution branch, a multihop graph rectify attention (MHRA) is designed to weight and correct the features extracted by graph convolution. In the convolutional branch, to solve the problem of dimensionality disaster caused by high spectral dimension with a small number of samples, a spectral intra group inter group feature extraction module (SI2FEM) based on spectral overlap grouping is constructed. In order to better fuse the features extracted from CNNs and GCNs, a Gaussian weighted fusion module (GWFM) is elaborately designed in this article. The features extracted by different branches are assigned different weights by GWFM through a 2-D Gaussian map and then fused. Numerous experiments were conducted on three common datasets and showed that the classification performance of the proposed MRSGFN is superior to other advanced methods.

The classification of land cover material based on hyperspectral image (HSI) has important research significance. Owing to the high cost of obtaining labeled samples and insufficient training samples, the research of cross-scene HSI classification (CS-HSIC) is receiving more and more attention. At present, the performance of the feature extraction module of CS-HSIC is relatively poor, and the number of training parameters is usually large. To fill the shortcomings of domain generalization (DG) methods and reduce the number of parameters, we propose a lightweight DG network with an attention-assisted cascaded bottleneck (ACB), and it adopts a lightweight bottleneck and multiattention design. This model is adept at extracting domain invariant information contained in the source domain (SD), and it may be flexibly embedded into other models. The experimental results show that our network has good classification accuracy and DG ability when the number of training samples is a little small. As a feature extraction subnetwork, it can improve the performance of the original model or reduce the required resources. The code will be available at https://github.com/zhulongyu1234/ACB.

In recent years, transformer has attracted the attention of many researchers in the field of remote sensing due to its ability to model global information. However, it is difficult to extract local features such as textures and edges of images, thereby limiting the performance of transformer-based hyperspectral image classification (HSIC). Currently, most existing transformer models for HSIC improve their performance by combining the powerful feature extraction ability of convolution, which also introduces a large number of trainable parameters and increases model complexity. To address this issue, this article proposes a dual attention transformer for attention head interaction (DAHIT) for HSIC. First, a spatial local bias module (SLBM) was designed in the spatial branch, which introduces local priors to extract local features effectively without introducing numerous trainable parameters. Then, an attention head interaction module (AHIM) was proposed, which can make the interaction of information obtained by different attention heads. Finally, a diagonal mask multiscale dual attention module (DAM) was constructed in the spectral branch to enhance the attention to the correlation among different spectral bands through diagonal masks and then to extract features at different scales through feature vectors. Through a series of experiments, the proposed DAHIT is evaluated on four commonly used HSI datasets. The experimental results show that compared with other advanced methods, the proposed DAHIT method exhibits excellent classification performance, demonstrating the effectiveness of the proposed method in HSIC.

In recent years, convolutional neural networks (CNNs) have achieved great success in hyperspectral image (HSI) classification tasks. CNNs focus more on the local features of HSIs. The recently emerging Transformer network has shown great interest in the global features of HSIs. However, existing Transformer networks only consider single-scale feature extraction and do not combine the advantages of multiscale feature extraction and Transformer global feature extraction. To address this issue, this article proposes a dual-branch multiscale Transformer (DBMST) for HSI classification. First, a large-size spectral convolution kernel is utilized for the spectral dimension of the hyperspectral cube for downsampling feature extraction. Next, a channel shrink soft split module (CS3M) is proposed, which not only solves the problem of missing local information in large-scale tokens but also extracts shallow features and performs dimensionality reduction on channels. Then, considering the different dimensions of features extracted at different scales in two branches, a pooled activation fusion module (PAFM) is carefully designed. Finally, the proposed DBMST is evaluated on three commonly used HSI datasets. The experimental results show that DBMST achieves better classification performance compared to other advanced networks, demonstrating the effectiveness of the proposed method in HSI classification.

Subgrade is an essential factor for the service life of highways and railways, especially the sulfate attack of subgrade made of saline soil. Therefore, it is necessary to improve the transport property of subgrade. In this study, sulfate transport of cementitious materials-solidified saline soil was assessed. Firstly, factors affecting sulfate transport were characterized in natural soil. Furthermore, fly ash and lime were implemented to prepare a new kind of improved soil. By analyzing the microstructure development, influence of fly ash and lime on sulfate transport was investigated and the improved mix proportion was addressed. Finally, the last section focused on the mechanism for sulfate resistance of the soil. Raw materials, phase assemblage improvement, and pore structure refinement caused by filler effect and pozzolanic reaction effect of fly ash were the main reasons for the sulfate resistance increase. This study is helpful for the service life improvement of highways and railways.

Improving strength while retaining good ductility is crucial for expanding the application of 2219 Al alloy. In this study, refined microstructure, excellent strength and ductility were obtained for the 2219 Al alloy under the combined effects of ultrasonic melt treatment (UMT), hot rolling and T6 heat treatment (HRT6). The mean grain size declined from 664.2 μm to 194.9 μm for the as-cast 2219 Al alloy after 240s UMT, with a refining efficiency of 70.7 %. Meanwhile, the Cu content in Al matrix was increased by 41.7 %, and the area fraction of reticular eutectic structure was accordingly lessened by 64.5 %. The nucleation of θʹʹ/θʹ-Al2Cu phase was actuated owing to the increased Cu content in Al matrix, resulting in more dispersive θʹʹ/θʹ-Al2Cu precipitates in the HRT6 alloy with UMT. Besides, the recrystallization was encouraged because the boundaries of refined as-cast grains provided more favorable nucleation sites, and the increased dispersive θʹʹ/θʹ-Al2Cu precipitates would inhibits the grain boundary merging during HRT6. Thus, the average dimension of the recrystallized grains was decreased to the lowest value of 71.3 μm in the HRT6 alloy with 240s UMT. Meanwhile, the ultimate tensile strength (UTS), yield strength (YS) and elongation (EL) were enhanced to 456.2 MPa, 307.0 MPa and 16.7 %, and precipitation strengthening contributed the most to the YS enhancement. The improved ductility was mainly due to the increased deformation capacity induced by the refined grains and reduced stress concentration caused by the dispersive θʹʹ/θʹ-Al2Cu particles.

Fiber orientations are constrained by boundary effects (e.g. size effect), which play a significant role in the macroscopic rigidity of fiber-reinforced concrete (FRC). However, how to quantify the size effect on the fiber orientation distribution functions (ODFs) remain to be open issues. In this work, a simple method to quantify the sectional-size impact on fiber ODFs is proposed. By disassembling the fiber rotatable range and fiber location on the cross-section into several small facets and points, respectively, we draw out the fiber ODFs with different sectional sizes through superposition. Then, the calculation method of the fiber orientation factor based on the ODF is proposed, and the framework of applying an ODF to calculate the stiffness matrix is given. Our results revealed the influence of the cross-section size on the elastic properties of FRC, which can provide a sound guidance for the design and optimization of fiber-based cement composites.Highlights1. A practical method for discretizing the cross-section of fiber-reinforced concrete (FRC) into a point set is proposed.2. A quantitative interplay between boundary constraints and size effects on fiber ODFs is developed.3. The calculation framework of the fiber orientation factor and elastic modulus of the FRC based on the ODF is established.4. The difference between the effective elastic modulus and Poisson’s ratio of FRCs with different boundary constraints is discussed.

Current materials comprising suture anchors used to reconstruct ligament-bone junctions still have limitation in biocompatibility, degradability or mechanical properties. Magnesium alloys are potential bone implant materials, and Mg2+has been shown to promote ligament-bone healing. Here, we used Mg-2 wt.% Zn-0.5 wt.% Y-1 wt.% Nd-0.5 wt.% Zr (ZE21C) alloy and Ti6Al4V (TC4) alloy to prepare suture anchors to reconstruct the patellar ligament-tibia in SD rats. We studied the degradation behavior of the ZE21C suture anchor via in vitroand in vivoexperiments and assessed its reparative effect on the ligament-bone junction. In vitro, the ZE21C suture anchor degraded gradually, and calcium and phosphorus products accumulated on its surface during degradation. In vivo, the ZE21C suture anchor could maintain its mechanical integrity within 12 weeks of implantation in rats. The tail of the ZE21C suture anchor in high stress concentration degraded rapidly during the early implantation stage (0–4weeks), while bone healing accelerated the degradation of the anchor head in the late implantation stage (4–12weeks). Radiological, histological, and biomechanical assays indicated that the ZE21C suture anchor promoted bone healing above the suture anchor and fibrocartilaginous interface regeneration in the ligament-bone junction, leading to better biomechanical strength than the TC4 group. Hence, this study provides a basis for further research on the clinical application of degradable magnesium alloy suture anchors.

Saturated hydrogenation of aromatic amines is a crucial methodology for synthesizing alicyclic amines. The dispersion of Ru atoms occurred after chemical reduction treatment, which was constrained to B and N coordination, further promoting the Ru species’ uniform distribution through interaction with the B–N bond of boron nitride. Moreover, the Ru coordinated with h-BN and further formed the stable catalytic active center. By virtue of its excellent stability, the hydrogenation of MDA executed in the fixed-bed reactor could run for more than 200 h. However, the density functional theory showed that the hydrogenation rate of the second benzene ring was comparatively slower because of the strong intermediate adsorption during the conversion from 4,4′-diaminodiphenylmethane to 4,4′-diaminodicyclohexylmethane. Furthermore, Ru/h-BN exhibited excellent performance in the hydrogenation of other aromatic amines, resulting in corresponding alicyclic amines. Therefore, this research offers a stable and effective catalyst for the green hydrogenation of aromatic amines leading to the production of alicyclic amines.

In this paper, citric acid-modified chitosan (CAMC) was added to cement slurry and its adsorption and hydration regulation mechanism on cement particles was studied. The rheological properties, agglomeration properties, and adsorption properties were characterized by different testing techniques. The effect of CAMC on the evolution of C-S-H gel and aluminum hydrate was further studied by the NMR technique. The results show that the negatively charged CAMC will adsorb on the surface of cement and agglomerate cement particles. In addition to being adsorbed on the positively charged aluminate phase surface under the electrostatic effect, CAMC was also adsorbed on the negatively charged silicate phase through complexation between carboxyl groups and calcium ions. The formation of complexes on the surface of the silicate phase hinders the hydration of calcium silicate (C3S).CAMC significantly reduced the hydration rate of early cement, hindering the nucleation of C-S-H and ettringite (AFt).

This paper proposes a novel remote sensing image segmentation method based on Gaussian mixture model and SURF algorithm. Firstly, Gaussian mixture model is used for remote sensing image segmentation. Then the surf matching algorithm is adopted for eliminating misidentified areas. The determinant of Hession matrix (DoH) is used to detect key points in the image. The non-maximum suppression method and interpolation operation are utilized to search and locate the extreme points. The maximum likelihood method is used to estimate model parameters. Some remote sensing images in THE DOTA data set are selected for experimental verification, and the results show that the new algorithm has obvious improvement in segmentation effect and efficiency. In the background complex image segmentation, the improved algorithm has more obvious advantages compared than state-of-the-art segmentation methods.

Membrane technology has been widely used in the field of water treatment, and the synthesis of membranes with better separation effects and better anti-pollution performances is also the focus of current research. Metal–organic framework materials (MOFs) have attracted extensive attention in many fields in recent years, especially in the synthesis of membranes with excellent water treatment and purification functions. Mixed matrix membranes (MMMs) prepared by adding MOFs to a polymer matrix can improve the performance of the membranes more effectively. In this paper, the advantages of representative MOFs and the preparation method of MOFs-based composite membranes are introduced. Compared with other traditional fillers, MOFs-based MMMs have more advantages in the separation process. It has been successfully integrated with the MOF membranes prepared, and various membrane processes are used in wastewater treatment or seawater desalination processes, for example, forward osmosis, ultrafiltration, nanofiltration, and reverse osmosis. At last, insights and viewpoints are put forward on research contents and development directions in the future. The membrane based on MOFs has excellent separation properties, low biological contamination, and high-water permeation flux, so it is expected to become the leading water treatment technology (in particular, desalination and wastewater treatment).

Real-time English Translation (RET) requires high network bandwidth and low network delay to provide better quality of experience, and even needs the support of massive connection to provide more services. For the three metrics, the traditional strategies are difficult to realize RET well. With the fast development of Mobile Edge Computing (MEC) and 5G network, the guarantee of three metrics has become very possible. Therefore, this paper studies MEC-based RET with 5G-driven network support, called 5GMR. On one hand, 5G-driven network has the natural properties to support high bandwidth, low delay and massive connection. On the other hand, MEC is used to offload the complex tasks related to the computation of English sentences into the edge server for the efficient computation, which not only saves energy consumption of mobile device but also decreases the whole network delay. In terms of the task scheduling in MEC, Genetic Algorithm (GA) is adopted to address it. The experimental results demonstrate that the proposed 5GMR is feasible and efficient.

In this study, citric acid-modified chitosan (CAMC) is prepared by an acylation reaction to regulate the exothermic process occurring in cement hydration. The thermodynamic process of cement hydration is characterized by isothermal calorimetry. The effects of various CAMC contents of 0%, 0.2%, 0.4%, and 0.6% on the hydration process were systematically studied by X-ray diffraction (XRD), inductively coupled plasma emission spectroscopy (ICP–OES), and proton nuclear magnetic resonance (1H-NMR). The results show that the adsorption effect delays the dissolution of alite (C3S). The content of CAMC is shown to have a significant effect on the formation time and amount of ettringite (AFt), and the competitive adsorption relationship increases the concentration of Ca2+. The increases in Ca2+and SO42-concentrations resulted in an increase in the AFt amount. With prolongation of the hydration, the capillary pores of the cement paste are transformed into gel pores, and the cement slurry exhibits a self-drying phenomenon.HighlightsDesigned and prepared citric acid modified chitosan.Adsorption delays the dissolution of C 3S and the formation of hydrated products.Competitive adsorption increases calcium ion concentration.CAMC delays the formation of AFt and increases the amount of AFt formation.

Hyperspectral images (HSIs) have been playing an important role in the field of ground object classification because of their rich spatial and spectral information. Aiming at how to extract complex feature information from HSIs, we propose a new feature fusion network based on dual attention (DAFFN) with dual attention mechanism, which is mainly used to capture more accurate global–local context attention features. The model extracts global context attention features using self-attention (SA) mechanism and local context attention features using cross-attention mechanism. Considering the problem that position information is easily lost during the conversion of attention mechanism, we propose a position self-calibration (PSC) module that can be flexibly embedded into two attention modules. In addition, to better integrate global and local features, we also designed a multiscale global and local feature fusion module (MSGL), which preserves more representative features with less communication costs by aggregating global and local attention features. We have carried out experiments on three commonly used hyperspectral datasets, and the classification results show that our model can achieve high classification accuracy even in the case of a limited number of samples.

Hyperspectral image (HSI) classification is one of the basic tasks of remote sensing image processing, which is to predict the label of each HSI pixel. Convolutional neural network (CNN) and graph convolution neural network (GCN) have become the current research focus due to their outstanding performance in the field of HSI classification in recent years. However, GCN is a transductive learning method, which needs all nodes to participate in the training process to get the node embedding. Graph sample and aggregate (GraphSAGE) is an important branch of graph neural network, which can flexibly aggregate new neighbor nodes in non-Euclidean data of any structure and capture long-range contextual relationships. Superpixel-based GraphSAGE can not only integrate the global spatial relationship of data but also further reduce its computing cost. CNN can extract pixel-level features in a small area, and our center attention module (CAM) and center weighted convolution (CW-Conv) can also improve the feature extraction ability of CNN by enhancing the dominant position of target pixels. In order to make full use of the advantages of CNN and GraphSAGE, we propose a center weighted convolution and GraphSAGE (CW-SAGE) cooperative network for HSI classification. Specifically, graph simple and aggregate branch is constructed by superpixel-based encoder and decoder modules, and then, pixel-level features are extracted by a central attention CNN. Finally, the features of the two branches are spliced together for feature fusion. We conduct experiments on three hyperspectral datasets and compare the results with other current state-of-the-art methods. A series of experiments demonstrates the advantages of our method.

In the past, deep learning (DL) technologies have been widely used in hyperspectral image (HSI) classification tasks. Among them, convolutional neural networks (CNNs) use fixed-size receptive field (RF) to obtain spectral and spatial features of HSIs, showing great feature extraction capabilities, which are one of the most popular DL frameworks. However, the convolution using local extraction and global parameter sharing mechanism pays more attention to spatial content information, which changes the spectral sequence information in the learned features. In addition, CNN is difficult to describe the long-distance correlation between HSI pixels and bands. To solve these problems, a spectral–spatial fusion Transformer network (S2FTNet) is proposed for the classification of HSIs. Specifically, S2FTNet adopts the Transformer framework to build a spatial Transformer module (SpaFormer) and a spectral Transformer module (SpeFormer) to capture image spatial and spectral long-distance dependencies. In addition, an adaptive spectral–spatial fusion mechanism (AS2FM) is proposed to effectively fuse the obtained advanced high-level semantic features. Finally, a large number of experiments were carried out on four datasets, Indian Pines, Pavia, Salinas, and WHU-Hi-LongKou, which verified that the proposed S2FTNet can provide better classification performance than other the state-of-the-art networks.

For hyperspectral image classification (HIC) tasks, most uncertainty-based active learning (AL) methods only consider the uncertainty, without considering the diversity of actively selected samples and the budget of expert labeling. In this article, we propose a collaborative AL (CAL) framework to address this problem. The proposed framework consists of two well-designed base classifiers and an ingenious CAL scheme that takes into account both the uncertainty and diversity of actively selected samples and the cost of expert annotation. Specifically, getting benefit from the ability of capsule networks (CapsNets) to accurately identify and locate features, we design two improved CapsNets. For these two networks, we call the first one Capsule Vision Transformer (CapsViT), which introduces vision Transformer (ViT) into the CapsNet to learn the global relationship between the capsule features. We call the second one Capsule GLOM (CapsGLOM); the basic structure of this network is derived from the GLOM system proposed by Geoffrey Hinton; and we learn from the way CapsNet constructs the primary capsules to improve its implementation details. CapsViT and CapsGLOM are used as the two base classifiers in the proposed CAL framework to select the most informative samples according to the CAL scheme under the premise of fully considering the cost of expert annotation. Experimental results on four benchmark hyperspectral image datasets show that our proposed CAL framework can achieve satisfactory classification results. At the same time, compared with other advanced deep models, our proposed CapsViT and CapsGLOM are also competitive in the supervised HIC tasks. The source code is available online (https://github.com/swiftest/CAL).

Autoencoder (AE) has been widely used in the field of hyperspectral anomaly detection. It is assumed that the background can be reconstructed well, but the anomalies cannot. Hence, the pixels with larger reconstruction errors are considered anomalies. However, owing to the strong nonlinear representation ability of AE, it is difficult to distinguish between background and anomalies. To address this problem, we propose a residual self-attention-based AE (RSAAE) for hyperspectral anomaly detection. RSAAE consists of dense residual self-attention (RSA) modules, an encoder, and a decoder. First, a novel RSA module is designed, which can effectively extract the main features and weaken the ability of subsequent networks to reconstruct anomalies, as well as preserve the original features to avoid the deterioration of network performance after the use of dense self-attention modules. Furthermore, inspired by manifold learning, we assume that the background is low-rank in the original space, and has the same property in the latent space after dimensionality reduction. We proposed a low-rank loss function to constrain the latent space, thereby suppressing anomaly reconstruction. Experiments on four real hyperspectral image (HSI) datasets showed that the proposed RSAAE method can produce more accurate detection results than eight popular methods.

Sub-pixel mapping (SPM) technology can analyze mixed pixels in spectral image and realize the transformation from abundance images to a fine sub-pixel classification image. Since the SPM belongs to an ill-posed issue, deviation information (DI) inevitably is in optimal abundance images. Due to simple structure and good robustness, most SPM approaches are according to the spatial dependence assumption; namely, the closer the spatial distance is, the more likely the subpixels belong to the same class. However, the existing SPM methods based on spatial dependence assumption cannot accurately measure the DI, which affects the accuracy of the final mapping result. To address this problem, the SPM based on DI measurement (DIM) is proposed in this work. The DIM uses a dual-scale spatial attraction model (DSAM) to process the coarse abundance images to obtain predicted abundance images. The fine prior images captured at different times from the same field of view are used to measure the deviation abundance images with the DI using a geographically weighted regression (GWR) model. The predicted abundance images and deviation abundance images are fused to derive optimal abundance images. Based on the proportion information on sub-pixels being classes in the optimal abundance images, the class labels are assigned to sub-pixels by label allocation method, yielding the final mapping result. The proposed DIM method is tested on National Land-cover Dataset, Rome Dataset, and Bastrop Fires dataset. The experimental results verify that the proposed DIM achieves the best performance with the overall accuracy of 97.26%, 88.15%, and 99.80% in the three experimental results.

In recent years, convolutional neural networks (CNNs) have been widely used in hyperspectral image classification (HSIC). However, the size of the convolutional kernel in CNNs is fixed, which makes it difficult to capture the dependence of long-range feature information. In addition, the extracted features often contain a large amount of redundant information. In order to alleviate these issues, a feature complementary attention network based on adaptive knowledge filtering (FCAN_AKF) is proposed in this article. First, in order to alleviate the problem that CNNs are difficult to capture the dependence between close-range and long-range spectral features due to the limited receptive field, a nonlocal band regrouping (NBR) strategy is designed. NBR enables CNN to capture nonlocal spectral features in a limited receptive field to establish the interdependence between close-range and long-range spectral features. In addition, the nonlocal features extracted after using NBR and the local features of the original hyperspectral image are integrated to achieve complementation between nonlocal features and local features. Then, in order to eliminate the interference of redundant information on the network, a dual-pyramid spectral–spatial attention (DPSSA) module is proposed and used to capture spectral–spatial attention. Next, an adaptive knowledge filter (AKF) is designed, which can adaptively further filter out redundant information and enhance feature information that is beneficial for classification. Finally, extensive experiments were conducted on three challenging datasets, demonstrating that the proposed method has stronger competitiveness compared to some state-of-the-art HSIC methods.

Poisson noise is one of the significant sources of noise present in hyperspectral imagery (HSI). In most of the existing denoising methods, Poisson noise is first transformed into Gaussian noise through the Anscombe transform and then removed. However, the use of Anscombe transform can give rise to transform errors that affect the final denoising results. In addition, blurs often contaminate the HSI during the imaging procedure, which makes it more difficult to remove the Poisson noise. In view of the above problems, under the maximum a posteriori (MAP) model, we propose a Poissonian blurred HSI denoising based on variable splitting and penalty technique (named as VSPT) to directly remove the Poissonian blurred HSI noise without using the Anscombe transform. By finding the minimum value of the negative logarithmic Poisson log-likelihood combined with the total variation (TV), the proposed method transforms the problem into two subproblems, which are easier to solve: 1) a TV regularized deconvolution problem and 2) an ordinary convex optimization problem. The experimental results show that the proposed VSPT method can effectively remove Poisson noise in HSI contaminated by blurs during the imaging procedure.

In recent years, pixel-wise hyperspectral image (HSI) classification has received growing attention in the field of remote sensing. Plenty of spectral–spatial convolutional neural network (CNN) methods with diverse attention mechanisms have been proposed for HSI classification due to the attention mechanisms being able to provide more flexibility over standard convolutional blocks. However, it remains a challenge to effectively extract multiscale features of high-resolution HSI in a real-world complex environment. In this article, we propose a pyramidal multiscale spectral–spatial convolutional network with polarized self-attention for pixel-wise HSI classification. It contains three stages: channel-wise feature extraction network, spatial-wise feature extraction network, and classification network, which are used to extract spectral features, extract spatial features, and generate classification results, respectively. Pyramidal convolutional blocks and polarized attention blocks are combined to extract spectral and spatial features of HSI. Furthermore, residual aggregation and one-shot aggregation are employed to better converge the network. The experimental results on several public HSI datasets demonstrate that the proposed network outperforms other related methods.

The emergence of convolutional neural networks (CNNs) has greatly promoted the development of hyperspectral image classification (HSIC). However, some serious problems are the lack of label samples in hyperspectral images (HSIs), and the spectral characteristics of different objects in HSIs are sometimes similar among classes. These problems hinder the improvement of HSIC performance. To this end, in this article, a positive feedback spatial-spectral correlation network based on spectral interclass slicing (PFSSC_SICS) is proposed. First, a spectral interclass slicing (SICS) strategy is designed, which can remove similar spectral signature between classes and reduce the impact of similar spectral signature of different classes on HSIC performance. Second, in order to solve the impact of the lack of labeled samples on HSIC, a positive feedback (PF) mechanism and a spatial-spectral correlation (SSC) module are introduced to extract deeper and more features. Finally, the experimental results show that the classification performance of the PFSSC_SICS is far exceed than that of some state-of-the-art methods.

Graphical abstract:

The problem of boreholes deflection in coal mines is common, especially in soft and hard composite coal seams. As a result, the distribution of boreholes is uneven and forms a blank zone in the drainage area, which fails to achieve the regional pre-draining, causing gas explosion or coal and gas outburst accidents. Aiming at the above problems, this paper developed the technology of anti-inclination for drilling based on the causes of borehole deflection in detail. The anti-inclination advice integrates straight keeping, slag removal and smooth back drilling with reasonable technical parameters. Full hole drilling was adopted to realize straight drilling. Three spiral grooving, vertical grooving and three-edge short-circuit were adopted for collaborative slag removal. Three alloy pieces were welded at the rear end of anti-inclination device to act as reverse drill bit, realizing smooth drilling withdrawal. For sticking problems, we improved the anti-inclination device. Three 20 mm millimeters wide vertical grooves were milled in each 120° direction around the device. After using the improved anti-inclination device, the average deviation of dip and azimuth was only 4.8 and 4.5 degrees, respectively, which were reduced by 51 and 44%. The borehole offset was reduced from 9.9 m to 2.9 m, a decrease of 71%, which improved the gas drainage effect and ensured the safe production of the mine.

To monitor, diagnose, and suppress the inter-turn short-circuit faults (ITSCFs) of a submersible motor, an approach for the on-line identification of its winding faults has been proposed based on monitoring the stator current. First, an ITSCF model with the global leakage referred to the stator is given. With this model, the detection parameter, which is equal to the ratio of the turns of the fault windings to the total turns of the windings in the healthy phase, can be derived. Second, a faulty model described by the fourth order state-space equation of the motor with a winding fault has been given. Based on sampled stator voltage and stator current, the detection parameter has been solved and used to estimate the location and the turns of inter-turn short-circuit windings of the motor in real time. The accuracy and the robustness of the proposed approach has been illustrated with a 1.5 kW motor that is fed by a 10 kW inverter. Experiment shows that the identification accuracy in terms of the number of the ITSCF windings of the motor stator is less than 3. It can give a reference for the on-line diagnose the ITSCFs of the stator windings of a submersible motor that works in 2 km deep well.

Studying the mechanism by which organic admixtures affect the hydration and dissolution of tricalcium silicate (C3S) reveals the effect of organic admixtures on ordinary Portland cement and enables target regulation of cement-based materials. In this study, the effects of citric-acid-modified chitosan (CAMC) on the hydration exotherm, hydration products, and microscopic morphology of C3S were investigated. The interface structures, ion adsorptions, and dissolution properties of CAMC and C3S were analyzed using a molecular dynamics simulation method. The results showed the presence of an attraction between the C3S surface and CAMC due to the existence of Op-Hw, Op-Cas, and Hp-Owion pairs. CAMC adsorbed most of the Ca ions released upon dissolution of C3S in the aqueous solution and the resulting pairs exhibited low solubilities. The Ca ions were located on the surfaces of C3S particles, preventing the dissolution of the particles and proving the interference effect of additives on the hydration of the cement silicate phase.