Micro silicon(mSi)is a promising anode candidate for all-solid-state batteries due to its high specific capacity,low side reactions,and high tap density.However,silicon suffers from its poor electronic and ionic condu...Micro silicon(mSi)is a promising anode candidate for all-solid-state batteries due to its high specific capacity,low side reactions,and high tap density.However,silicon suffers from its poor electronic and ionic conductivity,which is particularly severe on a micro scale and in solid-state systems,leading to increased polarization and inferior electrochemical performance.Doping can broaden the transmission pathways and reduce the diffusion energy barrier for electrons and lithium ions.However,achieving effective,uniform doping in mSi is challenging due to its longer diffusion paths and higher energy barriers.Therefore,current doping research is primarily limited to nanosilicon.In this study,we successfully used a Joule-heating activated staged thermal treatment to achieve full-depth doping of germanium(Ge)in the mSi substrate.The Joule-heating process activated the mSi substrate,resulting in abundant vacancy defects that reduced the diffusion barrier of Ge into the silicon lattice and facilitated full-depth Ge doping.Surprisingly,the resulting Si-Ge anode exhibited significantly enhanced electrical conductivity(70 times).Meanwhile,the improved Li-ion conductivity in mSi and the reduced Young’s modulus enhance the electrode reaction kinetics and integrity after cycling.Ge-doped silicon anodes demonstrate excellent electrochemical performance when applied in sulfide solid-state half-cells and full-cells.This work provides substantial insights into the rational structural design of mSi alloyed anode materials,paving the way for the development of high-performance solid-state Li-ion batteries.展开更多
Lithium?ion batteries(LIBs), which are high?energy?density and low?safety?risk secondary batteries, are underpinned to the rise in electrochemical energy storage devices that satisfy the urgent demands of the global e...Lithium?ion batteries(LIBs), which are high?energy?density and low?safety?risk secondary batteries, are underpinned to the rise in electrochemical energy storage devices that satisfy the urgent demands of the global energy storage market. With the aim of achiev?ing high energy density and fast?charging performance, the exploitation of simple and low?cost approaches for the production of high capacity, high density, high mass loading, and kinetically ion?accessible electrodes that maximize charge storage and transport in LIBs, is a critical need. Toward the construction of high?performance electrodes, carbons are promisingly used in the enhanced roles of active materials, electrochemi?cal reaction frameworks for high?capacity noncarbons, and lightweight current collectors. Here, we review recent advances in the carbon engi?neering of electrodes for excellent electrochemical performance and structural stability, which is enabled by assembled carbon architectures that guarantee su cient charge delivery and volume fluctuation bu ering inside the electrode during cycling. Some specific feasible assem?bly methods, synergism between structural design components of carbon assemblies, and electrochemical performance enhancement are highlighted. The precise design of carbon cages by the assembly of graphene units is potentially useful for the controlled preparation of high?capacity carbon?caged noncarbon anodes with volumetric capacities over 2100 mAh cm^(-3). Finally, insights are given on the prospects and challenges for designing carbon architectures for practical LIBs that simultaneously provide high energy densities(both gravimetric and volumetric) and high rate performance.展开更多
The flotation separation of magnesite from calcium-containing minerals has always been a difficult subject in minerals processing.This work studied the inhibition effects of carboxymethyl cellulose(CMC),sodium lignosu...The flotation separation of magnesite from calcium-containing minerals has always been a difficult subject in minerals processing.This work studied the inhibition effects of carboxymethyl cellulose(CMC),sodium lignosulphonate,polyaspartic acid(PASP)and sodium silicate on flotation behaviors of magnesite,dolomite and calcite,providing guidance for the development of reagents in magnesite flotation.The micro-flotation results showed that among these four depressants,sodium silicate presented the strongest selectivity due to the highest recovery difference,and the flotation separation of magnesite from dolomite and calcite could be achieved by using sodium silicate as the depressant.Contact angle measurement indicated that the addition of sodium silicate caused the largest differences in surface wettability of the three minerals,which was in line with micro-flotation tests.Furthermore,zeta potential test,the Fourier transform infrared(FT-IR)spectroscopy and atomic force microscope(AFM)imaging were used to reveal the inhibition mechanism of sodium silicate.The results indicated that the dominated component SiO(OH)3of sodium silicate could adsorb on minerals surfaces,and the adsorption of sodium silicate hardly affected the adsorption of NaOL on magnesite surface,but caused the reduction of NaOL adsorption on dolomite and calcite surfaces,thereby increasing the flotation selectivity.展开更多
Recently,video object segmentation has received great attention in the computer vision community.Most of the existing methods heavily rely on the pixel-wise human annotations,which are expensive and time-consuming to ...Recently,video object segmentation has received great attention in the computer vision community.Most of the existing methods heavily rely on the pixel-wise human annotations,which are expensive and time-consuming to obtain.To tackle this problem,we make an early attempt to achieve video object segmentation with scribble-level supervision,which can alleviate large amounts of human labor for collecting the manual annotation.However,using conventional network architectures and learning objective functions under this scenario cannot work well as the supervision information is highly sparse and incomplete.To address this issue,this paper introduces two novel elements to learn the video object segmentation model.The first one is the scribble attention module,which captures more accurate context information and learns an effective attention map to enhance the contrast between foreground and background.The other one is the scribble-supervised loss,which can optimize the unlabeled pixels and dynamically correct inaccurate segmented areas during the training stage.To evaluate the proposed method,we implement experiments on two video object segmentation benchmark datasets,You Tube-video object segmentation(VOS),and densely annotated video segmentation(DAVIS)-2017.We first generate the scribble annotations from the original per-pixel annotations.Then,we train our model and compare its test performance with the baseline models and other existing works.Extensive experiments demonstrate that the proposed method can work effectively and approach to the methods requiring the dense per-pixel annotations.展开更多
Segmenting the semantic regions of point clouds is a crucial step for intelligent agents to understand 3D scenes.Weakly supervised point cloud segmentation is highly desirable because entirely labelling point clouds i...Segmenting the semantic regions of point clouds is a crucial step for intelligent agents to understand 3D scenes.Weakly supervised point cloud segmentation is highly desirable because entirely labelling point clouds is highly time-consuming and costly.For the low-costing labelling of 3D point clouds,the scene-level label is one of the most effortless label strategies.However,due to the limitation of classifier discriminative capability and the orderless and structurless nature of the point cloud data,existing scene-level method is hard to transfer the semantic information,which usually leads to the under-activated or over-activated issues.To this end,a local semantic embedding network is introduced to learn local structural patterns and semantic propagation.Specifically,the proposed network contains graph convolution-based dilation and erosion embedding modules to implement‘inside-out’and‘outside-in’semantic information dissemination pathways.Therefore,the proposed weakly supervised learning framework could achieve the mutual propagation of semantic information in the foreground and background.Comprehensive experiments on the widely used ScanNet benchmark demonstrate the superior capacity of the proposed approach when compared to the current alternatives and baseline models.展开更多
A manipulator-type docking hardware-in-the-loop(HIL)simulation system is proposed in this paper,with further development of the space docking technology and corresponding requirements of the engineering project.First,...A manipulator-type docking hardware-in-the-loop(HIL)simulation system is proposed in this paper,with further development of the space docking technology and corresponding requirements of the engineering project.First,the structure of the manipulator-type HIL simulation system is explained.The mass and the flexibility of the manipulator has an important influence on the stability of the HIL system,which is the premise of accurately simulating actual space docking.Thus,the docking HIL simulation models of rigid,flexible and flexible-light space manipulators are established.The characteristics of the three HIL systems are studied from three important aspects:the system parameter configuration relation,the system stability condition and the dynamics frequency simulation ability.The key conclusions obtained were that the system satisfies stability or reproduction accuracy.Meanwhile,the influence of different manipulators on the system stability is further analyzed.The accuracy of the calculated results is verified experimentally.展开更多
A comprehensive method to evaluate the factors affecting the production capacity of horizontal wells in Carboniferous volcanic rocks after fracturing is investigated.A systematic approach combining gray correlation an...A comprehensive method to evaluate the factors affecting the production capacity of horizontal wells in Carboniferous volcanic rocks after fracturing is investigated.A systematic approach combining gray correlation analysis,hierarchical analysis and fuzzy evaluation is proposed.In particular,first the incidence of reservoir properties and fracturing parameters on production capacity is assessed.These parameters include reservoir base geological parameters(porosity,permeability,oil saturation,waterproof height)as well as engineering parameters(fracture halflength,fracture height,fracture conductivity,fracture distance).Afterwards,a two-by-two comparison judgment matrix of sensitive parameters is constructed by means of hierarchical analysis,and the weighting coefficients of the factors are determined,where oil saturation,fracture conductivity and fracture half-length are weighted higher.Finally,the horizontal wells in the target block are categorized in terms of production capacity based on the fuzzy evaluation method,and split accordingly into high-producing,relatively high-producing,medium-producing and low-producing wells.Such a categorization is intended to provide parametric guidance for reservoir fracturing and modification.展开更多
This study addresses the limitations of current aerial water-dropping models,which primarily focus on single-aircraft operations,with limited research on multi-aircraft firefighting strategies that could significantly...This study addresses the limitations of current aerial water-dropping models,which primarily focus on single-aircraft operations,with limited research on multi-aircraft firefighting strategies that could significantly enhance firefighting efficiency.Additionally,current model calibration techniques often rely on the real testing of firefighting,leading to a high cost of implementation.To overcome these challenges,an effective fire extinguishing model based on turbulent jet principles is developed to improve the efficiency of multi-aircraft firefighting,and an analytical solution is derived to the parameter optimization.To reduce the cost of the model calibration,a cost-effective calibration method is proposed without the requirement of setting a real fire in the grassland based on image processing techniques,which enables the determination of model parameters at reduced operational expenses.Furthermore,simulations and practical experiments validate the effectiveness of both the model and the calibration method,demonstrating substantial benefits in improving aerial firefighting strategies.展开更多
Increasing the density and thickness of electrodes is required to maximize the volumetric energy density of lithium-ion batteries for practical applications.However,dense and thick electrodes,especially highmass-conte...Increasing the density and thickness of electrodes is required to maximize the volumetric energy density of lithium-ion batteries for practical applications.However,dense and thick electrodes,especially highmass-content(>50 wt%) silicon anodes,have poor mechanical stability due to the presence of a large number of unstable interfaces between the silicon and conducting components during cycling.Here we report a network of mechanically robust carbon cages produced by the capillary shrinkage of graphene hydrogels that can contain the silicon nanoparticles in the cages and stabilize the silicon/carbon interfaces.In situ transmission electron microscope characterizations including compression and tearing of the structure and lithiation-induced silicon expansion experiments,have provided insight into the excellent confinement and buffering ability of this interface-strengthened graphene-caged silicon nanoparticle anode material.Consequently,a dense and thick silicon anode with reduced thickness fluctuations has been shown to deliver both high volumetric(>1000 mAh cm^-3) and areal(>6 mAh cm^-2)capacities together with excellent cycling capability.展开更多
Automatic and robust object detection in remote sensing images is of vital significance in real-world applications such as land resource management and disaster rescue.However,poor performance arises when the state-of...Automatic and robust object detection in remote sensing images is of vital significance in real-world applications such as land resource management and disaster rescue.However,poor performance arises when the state-of-the-art natural image detection algorithms are directly applied to remote sensing images,which largely results from the variations in object scale,aspect ratio,indistinguishable object appearances,and complex background scenario.In this paper,we propose a novel Feature Enhancement Network(FENet)for object detection in optical remote sensing images,which consists of a Dual Attention Feature Enhancement(DAFE)module and a Context Feature Enhancement(CFE)module.Specifically,the DAFE module is introduced to highlight the network to focus on the distinctive features of the objects of interest and suppress useless ones by jointly recalibrating the spatial and channel feature responses.The CFE module is designed to capture global context cues and selectively strengthen class-aware features by leveraging image-level contextual information that indicates the presence or absence of the object classes.To this end,we employ a context encoding loss to regularize the model training which promotes the object detector to understand the scene better and narrows the probable object categories in prediction.We achieve our proposed FENet by unifying DAFE and CFE into the framework of Faster R-CNN.In the experiments,we evaluate our proposed method on two large-scale remote sensing image object detection datasets including DIOR and DOTA and demonstrate its effectiveness compared with the baseline methods.展开更多
1 Introduction Endoscopy plays a crucial role in the diagnoses and treatment of gastrointestinal(GI)diseases[1],as it helps to identify abnormalities,classify lesion,and determine treatment methods.During GI endoscopi...1 Introduction Endoscopy plays a crucial role in the diagnoses and treatment of gastrointestinal(GI)diseases[1],as it helps to identify abnormalities,classify lesion,and determine treatment methods.During GI endoscopic examinations,physicians may encounter practical hindrances,i.e.,fatigue,stress,or limited experience,which can lead to erroneous results.Artificial intelligence(AI)-assisted GI endoscopy technology has emerged to address these limitations[2].展开更多
Direct hydrazine‐hydrogen peroxide fuel cells(DHzHPFCs)offer unique advantages for air‐independent applications,but their commercialization is impeded by the lack of high‐performance and low‐cost catalysts.This st...Direct hydrazine‐hydrogen peroxide fuel cells(DHzHPFCs)offer unique advantages for air‐independent applications,but their commercialization is impeded by the lack of high‐performance and low‐cost catalysts.This study reports a novel dual‐site Co‐Zn catalyst designed to enhance the hydrazine oxidation reaction(HzOR)activity.Density functional theory calculations suggested that incorporating Zn into Co catalysts can weaken the binding strength of the crucial N_(2)H_(3)*intermediate,which limits the ratedetermining N_(2)H_(3)*desorption step.The synthesized p‐Co_(9)Zn1 catalyst exhibited a remarkably low reaction potential of−0.15 V versus RHE at 10mAcm−2,outperforming monometallic Co catalysts.Experimental and computational analyses revealed dual active sites at the Co/ZnO interface,which facilitate N_(2)H_(3)*desorption and subsequent N_(2)H_(2)*formation.A liquidN_(2)H_(4)‐H_(2)O_(2)fuel cell with p‐Co_(9)Zn1 catalyst achieved a high open circuit voltage of 1.916 V and a maximum power density of 195mWcm^(−2),demonstrating the potential application of the dual‐site Co‐Zn catalyst.This rational design strategy of tuning the N_(2)H_(3)*binding energy through bimetallic interactions provides a pathway for developing efficient and economical non‐precious metal electrocatalysts for DHzHPFCs.展开更多
The adsorption of rare earth elements(REEs)from wastewater is vital for environmental protection and resource utilization.Adsorbents with magnetic properties are easy to separate but incorporating magnetic particles c...The adsorption of rare earth elements(REEs)from wastewater is vital for environmental protection and resource utilization.Adsorbents with magnetic properties are easy to separate but incorporating magnetic particles can reduce adsorption capacity by decreasing the surface area or blocking active sites.Herein,an efficient magnetic adsorbent(i.e.,Fe_(3)O_(4)@PDAPEI),consisting of an Fe_(3)O_(4)core,a polydopamine(PDA)intermediate layer and a polyethylenimine(PEI)outer layer,was designed to extract Gd^(3+),Nd^(3+),Ho^(3+),and Y^(3+)from low-concentration solutions with adsorption capacities of 168.3,168.5,179.7,and 180.3 mg/g,respectively.The adsorption capacities exceed those of most reported magnetic REE adsorbents in the literature.The adsorption behavior could be fitted to the pseudo-second-order model,intraparticle diffusion model,and Langmuir model.Fe_(3)O_(4)@PDAPEI exhibited good reusability,with the adsorption capacity remaining above 90%of the initial value after five reuse cycles.In addition,despite the presence of competing ions(i.e.,Na+,Mg2+,and Al^(3+))in model wastewater,the adsorption capacity could be maintained above 100 mg/g for all four REEs.The adsorption mechanism was investigated via density functional theory calculations,zeta potential measurements,and surface force measurements via atomic force microscopy.REEs could adsorb on Fe_(3)O_(4)@PDAPEI through binding to primary amines and electrostatic interactions.This work presents a highly efficient magnetic adsorbent and evaluates the underlying interaction mechanism from both theoretical and experimental perspectives,shedding light on facile and efficient REE recovery in various engineering processes.展开更多
The practical application of lithium-sulfur batteries with a high energy density has been plagued by the poor cycling stability of the sulfur cathode, which is a result of the insulating nature of sulfur and the disso...The practical application of lithium-sulfur batteries with a high energy density has been plagued by the poor cycling stability of the sulfur cathode, which is a result of the insulating nature of sulfur and the dissolution of polysulfides. Much work has been done to construct nanostructured or doped carbon as a porous or polar host for promising sulfur cathodes, although restricting the polysulfide shuttle effect by improving the redox reaction kinetics is more attractive. Herein, we present a well-designed strategy by introducing graphitic carbon nitride （g-C3N4） into a three-dimensional hierarchical porous graphene assembly to achieve a synergistic combination of confinement and catalyzation of polysulfides. The porous g-CBN4 nanosheets in situ formed inside the graphene network afford a highly accessible surface to catalyze the transformation of polysulfides, and the hierarchical porous graphene-assembled carbon can function as a conductive network and provide appropriate space for g-C3N4 catalysis in the sulfur cathode. Thus, this hybrid can effectively improve sulfur utilization and block the dissolution of polysulfides, achieving excellent cycling performance for sulfur cathodes in lithium-sulfur batteries.展开更多
Conventional carbon materials cannot combine high density and high porosity,which are required in many applications,typically for energy storage under a limited space.A novel highly dense yet porous carbon has previou...Conventional carbon materials cannot combine high density and high porosity,which are required in many applications,typically for energy storage under a limited space.A novel highly dense yet porous carbon has previously been produced from a three-dimensional(3D)reduced graphene oxide(r-GO)hydrogel by evaporation-induced drying.Here the mechanism of such a network shrinkage in r-GO hydrogel is specifically illustrated by the use of water and 1,4-dioxane,which have a sole difference in surface tension.As a result,the surface tension of the evaporating solvent determines the capillary forces in the nanochannels,which causes shrinkage of the r-GO network.More promisingly,the selection of a solvent with a known surface tension can precisely tune the microstructure associated with the density and porosity of the resulting porous carbon,rendering the porous carbon materials great potential in practical devices with high volumetric performance.展开更多
基金financially supported by the National Key Research and Development Program(2022YFE0127400)the National Natural Science Foundation of China(52172040,52202041,and U23B2077)+1 种基金Taishan Scholar Project of Shandong Province(tsqn202211086,ts202208832,tsqnz20221118)the Fundamental Research Funds for the Central Universities(23CX06055A).
文摘Micro silicon(mSi)is a promising anode candidate for all-solid-state batteries due to its high specific capacity,low side reactions,and high tap density.However,silicon suffers from its poor electronic and ionic conductivity,which is particularly severe on a micro scale and in solid-state systems,leading to increased polarization and inferior electrochemical performance.Doping can broaden the transmission pathways and reduce the diffusion energy barrier for electrons and lithium ions.However,achieving effective,uniform doping in mSi is challenging due to its longer diffusion paths and higher energy barriers.Therefore,current doping research is primarily limited to nanosilicon.In this study,we successfully used a Joule-heating activated staged thermal treatment to achieve full-depth doping of germanium(Ge)in the mSi substrate.The Joule-heating process activated the mSi substrate,resulting in abundant vacancy defects that reduced the diffusion barrier of Ge into the silicon lattice and facilitated full-depth Ge doping.Surprisingly,the resulting Si-Ge anode exhibited significantly enhanced electrical conductivity(70 times).Meanwhile,the improved Li-ion conductivity in mSi and the reduced Young’s modulus enhance the electrode reaction kinetics and integrity after cycling.Ge-doped silicon anodes demonstrate excellent electrochemical performance when applied in sulfide solid-state half-cells and full-cells.This work provides substantial insights into the rational structural design of mSi alloyed anode materials,paving the way for the development of high-performance solid-state Li-ion batteries.
基金supported by the National Science Fund for Distinguished Young Scholars of China (No. 51525204)National Key Basic Research Program of China (2014CB932400)the National Natural Science Foundation of China (No. 51872195 and U1401243)
文摘Lithium?ion batteries(LIBs), which are high?energy?density and low?safety?risk secondary batteries, are underpinned to the rise in electrochemical energy storage devices that satisfy the urgent demands of the global energy storage market. With the aim of achiev?ing high energy density and fast?charging performance, the exploitation of simple and low?cost approaches for the production of high capacity, high density, high mass loading, and kinetically ion?accessible electrodes that maximize charge storage and transport in LIBs, is a critical need. Toward the construction of high?performance electrodes, carbons are promisingly used in the enhanced roles of active materials, electrochemi?cal reaction frameworks for high?capacity noncarbons, and lightweight current collectors. Here, we review recent advances in the carbon engi?neering of electrodes for excellent electrochemical performance and structural stability, which is enabled by assembled carbon architectures that guarantee su cient charge delivery and volume fluctuation bu ering inside the electrode during cycling. Some specific feasible assem?bly methods, synergism between structural design components of carbon assemblies, and electrochemical performance enhancement are highlighted. The precise design of carbon cages by the assembly of graphene units is potentially useful for the controlled preparation of high?capacity carbon?caged noncarbon anodes with volumetric capacities over 2100 mAh cm^(-3). Finally, insights are given on the prospects and challenges for designing carbon architectures for practical LIBs that simultaneously provide high energy densities(both gravimetric and volumetric) and high rate performance.
基金supported by the Youth Science Foundation of China(No.52004333)the Key Laboratory of Hunan Province for Clean and Efficiency Utilization of Strategic Calcium-containing Mineral Resources(No.2018TP1002).
文摘The flotation separation of magnesite from calcium-containing minerals has always been a difficult subject in minerals processing.This work studied the inhibition effects of carboxymethyl cellulose(CMC),sodium lignosulphonate,polyaspartic acid(PASP)and sodium silicate on flotation behaviors of magnesite,dolomite and calcite,providing guidance for the development of reagents in magnesite flotation.The micro-flotation results showed that among these four depressants,sodium silicate presented the strongest selectivity due to the highest recovery difference,and the flotation separation of magnesite from dolomite and calcite could be achieved by using sodium silicate as the depressant.Contact angle measurement indicated that the addition of sodium silicate caused the largest differences in surface wettability of the three minerals,which was in line with micro-flotation tests.Furthermore,zeta potential test,the Fourier transform infrared(FT-IR)spectroscopy and atomic force microscope(AFM)imaging were used to reveal the inhibition mechanism of sodium silicate.The results indicated that the dominated component SiO(OH)3of sodium silicate could adsorb on minerals surfaces,and the adsorption of sodium silicate hardly affected the adsorption of NaOL on magnesite surface,but caused the reduction of NaOL adsorption on dolomite and calcite surfaces,thereby increasing the flotation selectivity.
基金supported in part by the National Key R&D Program of China(2017YFB0502904)the National Science Foundation of China(61876140)。
文摘Recently,video object segmentation has received great attention in the computer vision community.Most of the existing methods heavily rely on the pixel-wise human annotations,which are expensive and time-consuming to obtain.To tackle this problem,we make an early attempt to achieve video object segmentation with scribble-level supervision,which can alleviate large amounts of human labor for collecting the manual annotation.However,using conventional network architectures and learning objective functions under this scenario cannot work well as the supervision information is highly sparse and incomplete.To address this issue,this paper introduces two novel elements to learn the video object segmentation model.The first one is the scribble attention module,which captures more accurate context information and learns an effective attention map to enhance the contrast between foreground and background.The other one is the scribble-supervised loss,which can optimize the unlabeled pixels and dynamically correct inaccurate segmented areas during the training stage.To evaluate the proposed method,we implement experiments on two video object segmentation benchmark datasets,You Tube-video object segmentation(VOS),and densely annotated video segmentation(DAVIS)-2017.We first generate the scribble annotations from the original per-pixel annotations.Then,we train our model and compare its test performance with the baseline models and other existing works.Extensive experiments demonstrate that the proposed method can work effectively and approach to the methods requiring the dense per-pixel annotations.
基金Key-Area Research and Development Program of Guangdong Province,Grant/Award Number:2021B0101200001National Natural Science Foundation of China,Grant/Award Numbers:61876140,U20B2065,U21B2048Open Research Projects of Zhejiang Lab,Grant/Award Number:2019KD0AD01/010。
文摘Segmenting the semantic regions of point clouds is a crucial step for intelligent agents to understand 3D scenes.Weakly supervised point cloud segmentation is highly desirable because entirely labelling point clouds is highly time-consuming and costly.For the low-costing labelling of 3D point clouds,the scene-level label is one of the most effortless label strategies.However,due to the limitation of classifier discriminative capability and the orderless and structurless nature of the point cloud data,existing scene-level method is hard to transfer the semantic information,which usually leads to the under-activated or over-activated issues.To this end,a local semantic embedding network is introduced to learn local structural patterns and semantic propagation.Specifically,the proposed network contains graph convolution-based dilation and erosion embedding modules to implement‘inside-out’and‘outside-in’semantic information dissemination pathways.Therefore,the proposed weakly supervised learning framework could achieve the mutual propagation of semantic information in the foreground and background.Comprehensive experiments on the widely used ScanNet benchmark demonstrate the superior capacity of the proposed approach when compared to the current alternatives and baseline models.
基金Supported by the National Natural Science Foundation of China(51475116)。
文摘A manipulator-type docking hardware-in-the-loop(HIL)simulation system is proposed in this paper,with further development of the space docking technology and corresponding requirements of the engineering project.First,the structure of the manipulator-type HIL simulation system is explained.The mass and the flexibility of the manipulator has an important influence on the stability of the HIL system,which is the premise of accurately simulating actual space docking.Thus,the docking HIL simulation models of rigid,flexible and flexible-light space manipulators are established.The characteristics of the three HIL systems are studied from three important aspects:the system parameter configuration relation,the system stability condition and the dynamics frequency simulation ability.The key conclusions obtained were that the system satisfies stability or reproduction accuracy.Meanwhile,the influence of different manipulators on the system stability is further analyzed.The accuracy of the calculated results is verified experimentally.
基金Natural Science Foundation of China(NSFC),Grant No.52404016(Lijuan Huang)Natural Science Foundation of Hubei Province,Grant No.2024AFB322(Lijuan Huang)Open Fund of Hubei Key Laboratory of Oil and Gas Drilling and Production Engineering(Yangtze University),Grant No.YQZC202405(Lijuan Huang).
文摘A comprehensive method to evaluate the factors affecting the production capacity of horizontal wells in Carboniferous volcanic rocks after fracturing is investigated.A systematic approach combining gray correlation analysis,hierarchical analysis and fuzzy evaluation is proposed.In particular,first the incidence of reservoir properties and fracturing parameters on production capacity is assessed.These parameters include reservoir base geological parameters(porosity,permeability,oil saturation,waterproof height)as well as engineering parameters(fracture halflength,fracture height,fracture conductivity,fracture distance).Afterwards,a two-by-two comparison judgment matrix of sensitive parameters is constructed by means of hierarchical analysis,and the weighting coefficients of the factors are determined,where oil saturation,fracture conductivity and fracture half-length are weighted higher.Finally,the horizontal wells in the target block are categorized in terms of production capacity based on the fuzzy evaluation method,and split accordingly into high-producing,relatively high-producing,medium-producing and low-producing wells.Such a categorization is intended to provide parametric guidance for reservoir fracturing and modification.
基金supported by the“Intelligent Multi-source Autonomous Navigation”Basic Science Center Programthe General Programs of National Natural Science Foundation of China(Grant Nos.62388101,52372434,52472439)Aeronautical Science Foundation of China(Grant No.2019ZA053008)。
文摘This study addresses the limitations of current aerial water-dropping models,which primarily focus on single-aircraft operations,with limited research on multi-aircraft firefighting strategies that could significantly enhance firefighting efficiency.Additionally,current model calibration techniques often rely on the real testing of firefighting,leading to a high cost of implementation.To overcome these challenges,an effective fire extinguishing model based on turbulent jet principles is developed to improve the efficiency of multi-aircraft firefighting,and an analytical solution is derived to the parameter optimization.To reduce the cost of the model calibration,a cost-effective calibration method is proposed without the requirement of setting a real fire in the grassland based on image processing techniques,which enables the determination of model parameters at reduced operational expenses.Furthermore,simulations and practical experiments validate the effectiveness of both the model and the calibration method,demonstrating substantial benefits in improving aerial firefighting strategies.
基金the National Natural Science Foundation of China(51872195)the National Science Fund for Distinguished Young Scholars of China(51525204)+1 种基金JSPS KAKENHI(20K05281)the Beijing Natural Science Foundation(2192061)。
文摘Increasing the density and thickness of electrodes is required to maximize the volumetric energy density of lithium-ion batteries for practical applications.However,dense and thick electrodes,especially highmass-content(>50 wt%) silicon anodes,have poor mechanical stability due to the presence of a large number of unstable interfaces between the silicon and conducting components during cycling.Here we report a network of mechanically robust carbon cages produced by the capillary shrinkage of graphene hydrogels that can contain the silicon nanoparticles in the cages and stabilize the silicon/carbon interfaces.In situ transmission electron microscope characterizations including compression and tearing of the structure and lithiation-induced silicon expansion experiments,have provided insight into the excellent confinement and buffering ability of this interface-strengthened graphene-caged silicon nanoparticle anode material.Consequently,a dense and thick silicon anode with reduced thickness fluctuations has been shown to deliver both high volumetric(>1000 mAh cm^-3) and areal(>6 mAh cm^-2)capacities together with excellent cycling capability.
基金supported in part by the National Science Foundation of China under Grant 61772425in part by the Shaanxi Science Foundation for Distinguished Young Scholars under Grant 2021JC-16.
文摘Automatic and robust object detection in remote sensing images is of vital significance in real-world applications such as land resource management and disaster rescue.However,poor performance arises when the state-of-the-art natural image detection algorithms are directly applied to remote sensing images,which largely results from the variations in object scale,aspect ratio,indistinguishable object appearances,and complex background scenario.In this paper,we propose a novel Feature Enhancement Network(FENet)for object detection in optical remote sensing images,which consists of a Dual Attention Feature Enhancement(DAFE)module and a Context Feature Enhancement(CFE)module.Specifically,the DAFE module is introduced to highlight the network to focus on the distinctive features of the objects of interest and suppress useless ones by jointly recalibrating the spatial and channel feature responses.The CFE module is designed to capture global context cues and selectively strengthen class-aware features by leveraging image-level contextual information that indicates the presence or absence of the object classes.To this end,we employ a context encoding loss to regularize the model training which promotes the object detector to understand the scene better and narrows the probable object categories in prediction.We achieve our proposed FENet by unifying DAFE and CFE into the framework of Faster R-CNN.In the experiments,we evaluate our proposed method on two large-scale remote sensing image object detection datasets including DIOR and DOTA and demonstrate its effectiveness compared with the baseline methods.
基金supported in part by the National Natural Science Foundation of China(Grant Nos.62272468,62003256,62027813,U1801265,62293543,62322605,62036005,62202015,and U21B2048)the Key-Area Research and Development Program of Shaanxi Province(2023-ZDLSF-41)+2 种基金the Anhui Medical University(2022xkj105,2023cy021)the Anhui Provincial Key R&D Program(2023s07020001)the University Synergy Innovation Program of Anhui Province(GXXT-2022-052).
文摘1 Introduction Endoscopy plays a crucial role in the diagnoses and treatment of gastrointestinal(GI)diseases[1],as it helps to identify abnormalities,classify lesion,and determine treatment methods.During GI endoscopic examinations,physicians may encounter practical hindrances,i.e.,fatigue,stress,or limited experience,which can lead to erroneous results.Artificial intelligence(AI)-assisted GI endoscopy technology has emerged to address these limitations[2].
基金support was provided by the National Key Research and Development Program of China,Grant/Award Number:2023YFB4203900the Fundamental Research Funds for the Central Universities,Grant/Award Number:226‐2024‐00075.
文摘Direct hydrazine‐hydrogen peroxide fuel cells(DHzHPFCs)offer unique advantages for air‐independent applications,but their commercialization is impeded by the lack of high‐performance and low‐cost catalysts.This study reports a novel dual‐site Co‐Zn catalyst designed to enhance the hydrazine oxidation reaction(HzOR)activity.Density functional theory calculations suggested that incorporating Zn into Co catalysts can weaken the binding strength of the crucial N_(2)H_(3)*intermediate,which limits the ratedetermining N_(2)H_(3)*desorption step.The synthesized p‐Co_(9)Zn1 catalyst exhibited a remarkably low reaction potential of−0.15 V versus RHE at 10mAcm−2,outperforming monometallic Co catalysts.Experimental and computational analyses revealed dual active sites at the Co/ZnO interface,which facilitate N_(2)H_(3)*desorption and subsequent N_(2)H_(2)*formation.A liquidN_(2)H_(4)‐H_(2)O_(2)fuel cell with p‐Co_(9)Zn1 catalyst achieved a high open circuit voltage of 1.916 V and a maximum power density of 195mWcm^(−2),demonstrating the potential application of the dual‐site Co‐Zn catalyst.This rational design strategy of tuning the N_(2)H_(3)*binding energy through bimetallic interactions provides a pathway for developing efficient and economical non‐precious metal electrocatalysts for DHzHPFCs.
基金supported by the Fundamental Research Funds for the Central Universities(2232024D-05,China)the National Natural Science Foundation of China(52374293,52174269)the Natural Sciences and Engineering Research Council of Canada(NSERC),the Canada Foundation for Innovation(CFI),and the Canada Research Chairs Program.
文摘The adsorption of rare earth elements(REEs)from wastewater is vital for environmental protection and resource utilization.Adsorbents with magnetic properties are easy to separate but incorporating magnetic particles can reduce adsorption capacity by decreasing the surface area or blocking active sites.Herein,an efficient magnetic adsorbent(i.e.,Fe_(3)O_(4)@PDAPEI),consisting of an Fe_(3)O_(4)core,a polydopamine(PDA)intermediate layer and a polyethylenimine(PEI)outer layer,was designed to extract Gd^(3+),Nd^(3+),Ho^(3+),and Y^(3+)from low-concentration solutions with adsorption capacities of 168.3,168.5,179.7,and 180.3 mg/g,respectively.The adsorption capacities exceed those of most reported magnetic REE adsorbents in the literature.The adsorption behavior could be fitted to the pseudo-second-order model,intraparticle diffusion model,and Langmuir model.Fe_(3)O_(4)@PDAPEI exhibited good reusability,with the adsorption capacity remaining above 90%of the initial value after five reuse cycles.In addition,despite the presence of competing ions(i.e.,Na+,Mg2+,and Al^(3+))in model wastewater,the adsorption capacity could be maintained above 100 mg/g for all four REEs.The adsorption mechanism was investigated via density functional theory calculations,zeta potential measurements,and surface force measurements via atomic force microscopy.REEs could adsorb on Fe_(3)O_(4)@PDAPEI through binding to primary amines and electrostatic interactions.This work presents a highly efficient magnetic adsorbent and evaluates the underlying interaction mechanism from both theoretical and experimental perspectives,shedding light on facile and efficient REE recovery in various engineering processes.
文摘The practical application of lithium-sulfur batteries with a high energy density has been plagued by the poor cycling stability of the sulfur cathode, which is a result of the insulating nature of sulfur and the dissolution of polysulfides. Much work has been done to construct nanostructured or doped carbon as a porous or polar host for promising sulfur cathodes, although restricting the polysulfide shuttle effect by improving the redox reaction kinetics is more attractive. Herein, we present a well-designed strategy by introducing graphitic carbon nitride （g-C3N4） into a three-dimensional hierarchical porous graphene assembly to achieve a synergistic combination of confinement and catalyzation of polysulfides. The porous g-CBN4 nanosheets in situ formed inside the graphene network afford a highly accessible surface to catalyze the transformation of polysulfides, and the hierarchical porous graphene-assembled carbon can function as a conductive network and provide appropriate space for g-C3N4 catalysis in the sulfur cathode. Thus, this hybrid can effectively improve sulfur utilization and block the dissolution of polysulfides, achieving excellent cycling performance for sulfur cathodes in lithium-sulfur batteries.
基金This work was supported by the National Natural Science Fund for the Distinguished Young Scholars,China(51525204)the National Natural Science Foundation of China(51702229 and 51872195)the CAS Key Laboratory of Carbon Materials(KLCM KFJJ1704).
文摘Conventional carbon materials cannot combine high density and high porosity,which are required in many applications,typically for energy storage under a limited space.A novel highly dense yet porous carbon has previously been produced from a three-dimensional(3D)reduced graphene oxide(r-GO)hydrogel by evaporation-induced drying.Here the mechanism of such a network shrinkage in r-GO hydrogel is specifically illustrated by the use of water and 1,4-dioxane,which have a sole difference in surface tension.As a result,the surface tension of the evaporating solvent determines the capillary forces in the nanochannels,which causes shrinkage of the r-GO network.More promisingly,the selection of a solvent with a known surface tension can precisely tune the microstructure associated with the density and porosity of the resulting porous carbon,rendering the porous carbon materials great potential in practical devices with high volumetric performance.
