Search results

The paper aims to tackle the classification of Remote Sensing Images (RSIs), which presents a significant challenge for computer algorithms due to the inherent characteristics of clustered ground objects and noisy backgrounds. Recent research typically leverages larger volume models to achieve advanced performance. However, the operating environments of remote sensing commonly cannot provide unconstrained computational and storage resources. It requires lightweight algorithms with exceptional generalization capabilities.

This study introduces an efficient knowledge distillation (KD) method to build a lightweight yet precise convolutional neural network (CNN) classifier. This method also aims to substantially decrease the training time expenses commonly linked with traditional KD techniques. This approach entails extensive alterations to both the model training framework and the distillation process, each tailored to the unique characteristics of RSIs. In particular, this study establishes a robust ensemble teacher by independently training two CNN models using a customized, efficient training algorithm. Following this, this study modifies a KD loss function to mitigate the suppression of non-target category predictions, which are essential for capturing the inter- and intra-similarity of RSIs.

This study validated the student model, termed KD-enhanced network (KDE-Net), obtained through the KD process on three benchmark RSI data sets. The KDE-Net surpasses 42 other state-of-the-art methods in the literature published from 2020 to 2023. Compared to the top-ranked method’s performance on the challenging NWPU45 data set, KDE-Net demonstrated a noticeable 0.4% increase in overall accuracy with a significant 88% reduction in parameters. Meanwhile, this study’s reformed KD framework significantly enhances the knowledge transfer speed by at least three times.

This study illustrates that the logit-based KD technique can effectively develop lightweight CNN classifiers for RSI classification without substantial sacrifices in computation and storage costs. Compared to neural architecture search or other methods aiming to provide lightweight solutions, this study’s KDE-Net, based on the inherent characteristics of RSIs, is currently more efficient in constructing accurate yet lightweight classifiers for RSI classification.

A chromate conversion coating was prepared on the surface of bare AA2024 aluminum alloy by direct immersion in the chromating treatment bath, and the corrosion behavior of chromated AA2024 aluminum alloy in 3.5 per cent NaCl solution was studied by electrochemical measurement and microstructural observation.

According to the polarization curve test and the scanning electron microscope observation, the corrosion evolution of chromated AA2024 in 3.5 per cent NaCl solution was divided into the following three stages: coating failure, pitting corrosion and intergranular corrosion (IGC).

In the first stage, the chromate coating degraded gradually due to the combined action of chloride anions and water molecules, resulting in the complete exposure of AA2024 substrate to 3.5 per cent NaCl solution. Subsequently, in the second stage, chloride anions adsorbed at the sites of θ phase (Al2Cu) and S phase (Al2CuMg) on the AA2024 surface preferentially, and some corrosion pits initiated at the above two sites and propagated towards the deep of crystal grains. However, the propagation of a pit terminated when the pit front arrived at the adjacent grain boundary, where the initiation of IGC occurred.

Finally, in the third stage, the corrosion proceeded along the continuous grain boundary net and penetrated the internal of AA2024 substrate, resulting in the propagation of IGC. The related corrosion mechanisms for the bare and the chromated AA2024 were also discussed.

The purpose of this work is to reveal the mechanism of WO₄²⁻ on surface passivation for Q235 carbon steel in tungstate solution.

In Na₂WO₄ solutions with the different concentrations of WO₄²⁻, the spontaneous passivation occurred on the surface of Q235 carbon steel when the concentration of WO₄²⁻ was up to 0.13 mmol/L, which was attributed to the formations of the inner deposition film and the outer adsorption film on the Q235 surface under the action of WO₄²⁻.

The inner deposition film presented a two-layer microstructure: the inside layer was composed of Fe₂O₃ mainly, and the outside layer comprised Fe(OH)2•nH₂O, Fe(OH)3•nH₂O, FeWO₄ and Fe₂(WO₄)3.

Both FeWO₄ and Fe₂(WO₄)3 repaired the defects in the outside layer of the inner deposition film; however, the outer adsorption film played a more important role in the surface passivation than the inner deposition film did.

The aim of this work was to propose a method to prepare composite phosphate conversion coating (CPCC), including ternary phosphate conversion coating (TPCC) and binary phosphate conversion coatings (BPCC), with one-step chemical conversion and to reveal and compare the corrosion resistance between TPCC and BPCC.

In this work, a calcium–manganese–zinc (Ca–Mn–Zn) TPCC was prepared on the surface of magnesium alloy (MA) AZ91D with one-step chemical conversion method; for Ca-Mn-Zn@TPCC, its microstructure was characterized with scanning electron microscope observation and scanning tunneling microscope detection, and its composition was characterized with energy dispersion spectroscopy and X-ray photoelectron spectroscopy analyses. Particularly, the corrosion resistance of Ca-Mn-Zn@TPCC and its comparison with Ca–Mn, Ca–Zn and Mn–Zn BPCCs were clarified with electrochemical and immersion measurements.

Ca-Mn-Zn@TPCC, which was composed of Ca, Mn, Zn, P and O, exhibited a mud-shaped with cracks microstructure, and the average crack width, terrain fluctuation and coating thickness were 0.61 µm, 23.78 nm and 2.47 µm, respectively. Ca-Mn-Zn@TPCC provided good corrosion resistance to MA AZ91D; in NaCl solution, the total degradation of Ca-Mn-Zn@TPCC consumed eight days; corrosion products with poor adhesion peeled out from Ca-Mn-Zn@TPCC-coated MA AZ91D spontaneously. Besides, the corrosion resistance of Ca-Mn-Zn@TPCC was better than that of Ca-Mn@BPCC, Ca-Zn@BPCC or Mn-Zn@BPCC.

The successful preparation of Ca-Mn-Zn@TPCC on MA AZ91D surface confirmed the proposed method to prepare CPCC with one-step chemical conversion was feasible; at the same time, it was further confirmed that for phosphate conversion coating, ternary coating had better corrosion resistance than binary coating did.

The purpose of this paper is to reveal the mechanism of nitrite (NO_2⁻) for the surface passivation of carbon steels in acidic environments through investigating the influences of 0.01 mol/L NaNO₂ addition on the corrosion and passivation behaviors of Q235 carbon steel in acidic phosphate buffer (APB) solutions (pH 2 to 6).

The electrochemical techniques including open circle potential evolution, potentiodynamic polarization, electrochemical impedance spectroscopy and cyclic voltammetry were applied.

In APB solutions without NO_2⁻, the Q235 steel presented the electrochemical behaviors of activation (A), activation-passivation-transpassivation and self-passivation-transpassivation at pH 2 to 4, pH 5 and pH 6, respectively; the corrosion rate decreased with the up of pH value, and the surface passivation occurred in the pH 5 and pH 6 solutions only: the anodic passivation at pH 5 and the spontaneous passivation at pH 6.

In APB solutions without NO_2⁻, the corrosion rate decreased with the up of pH value, and the surface passivation occurred in the pH 5 and pH 6 solutions only: the anodic passivation at pH 5 and the spontaneous passivation at pH 6. With the addition of 0.01 mol/L NaNO₂, into APB solutions, the variation of corrosion rate showed the same rule, but the surface passivation occurred over the whole acidic pH range, including the anodic passivation at pH 2 to 4 and the spontaneous passivation at pH 5 to 6.

This paper aims to investigate magnetic field anneal in micro-patterned Co-based amorphous ribbon on giant magneto-impedance (GMI) effect enhancement.

The amorphous ribbons were annealed in transverse and longitudinal magnetic field. The influence of different field annealing directions on GMI effect and impedance Z, resistance R and reactance X with a series of line width have been deeply analyzed.

In comparison with GMI sensors microfabricated by unannealed and transversal field annealed ribbons, GMI sensor which was designed and microfabricated by longitudinal field anneal ribbon performs better. The results can be explained by the domain wall motion and domain rotation during annealing process and the geometric structure of Co-based GMI sensor. In addition, shrinking the line width of GMI sensor can promote GMI effect significantly because of the effect of demagnetizing field, and the optimum GMI ratio is 209.7 per cent in longitudinal field annealed GMI sensor with 200 μm line width.

In conclusion, annealing in longitudinal magnetic field and decreasing line width can enhance GMI effect in micro-patterned Co-based amorphous ribbon.

Drawing on job crafting theory and the social cognitive theory of work satisfaction, the purpose of this study is to examine the impacts of task, relational and cognitive crafting on job satisfaction and how occupational self-efficacy (OSE) and hierarchical organizational culture affect the execution and outcome of job crafting behaviors.

Data were collected at four time points among Chinese media professionals (N = 198) during three consecutive months. Regression analysis and the bootstrap method were used for hypothesis testing.

Only cognitive crafting was found to mediate the positive relationship between OSE and job satisfaction. In a hierarchical organizational culture, the positive effects of OSE on task, relational and cognitive crafting became stronger, while the effect of task crafting on job satisfaction became negative. The authors also identified a conditional negative effect of OSE on job satisfaction via task crafting. The mediating effect of cognitive crafting was consistent despite the levels of hierarchical culture.

To retain new hires, organizations should look for efficacious individuals and encourage the individuals to cognitively craft individuals' jobs, especially in a hierarchical culture. Moreover, job crafting interventions should be carefully designed in this type of culture.

This research identifies the three dimensions of job crafting as mediators that link OSE to job satisfaction and reveals the role of hierarchical culture in moderating these relationships.

Studies in hospitality and tourism have seldom investigated the role of employee commitment to building customers' attitudinal loyalty. This study examines the impact of employee commitment on customers' attitude-based loyalty. The study contributes to knowledge of how employees' affective attitude (i.e. employee commitment) impacts and mediates the relationships within this model by considering service quality attributes separately in the context of China's Greater Bay Area.

Using a quantitative approach, 664 customers visiting hotels and tourist attractions within three cities of Hong Kong, Macau and Zhuhai were surveyed. A convenience sampling technique was employed to administer questionnaires within these contexts. A structural equation modeling (SEM) using AMOS software was used to test the relationships in the proposed model.

The results suggest that while service quality attributes have a different impact on employee commitment, employee commitment plays a response-predictor-mediator role in the attitudinal loyalty framework. For instance, personal interactions and technical quality are significant predictors of employee commitment. Employee commitment influences customer satisfaction and behavioral intentions. Moreover, employee commitment fully mediates the relationship between technical quality and customer satisfaction and partially mediates the association between personal interaction and customer satisfaction.

Given that employee commitment could be derived from personal interaction with customers, hoteliers and destination management organizations should encourage customers through their websites to be responsive to employees by providing constructive feedback on their service delivery. Management of hotels, attractions and destinations need to motivate employees through incentives such as pay raise, bonuses, time-off and paid holidays.

The paper is inimitable in its attempt to extend the customer attitudinal loyalty debate by including employees' attitude (i.e. commitment) in the measurement of customers' attitudinal loyalty in the hospitality and tourism industry.

Science and technology innovation policy has important strategic significance with respect to the promotion of an innovation orientation in our country, and the classification and measurement of regional science and technology innovation policy urgently require research attention.

In this paper, we use text mining and principal component analysis to analyze the classification and measurement of technology innovation policy based on data obtained from Zhongguancun Science Park.

The empirical results indicate that regional science and technology innovation policy can be divided into four types: authoritative, guiding, urgent and periodical. The key measurements are function type, intensity, resource supply, funding level and funding effectiveness.

A comparative analysis is performed to investigate the different types of regional science and technology innovation policy measurement. Additionally, the study’s limitations are discussed, and future research directions are proposed.

The purpose of this paper is to discuss an adaptive SLA mechanism for service sharing in virtual environment, which can organize and govern QoS items in terms of service execution time, reliability, and availability, and provides a common understanding about services, responsibilities, priorities, guarantees and warranties-related virtual cooperative issues.

The management framework for SLA is introduced, based on which the whole process including SLA contract, adaptive SLA negotiation strategy, SLA deployment and SLA assessment are discussed, and the prototype is implemented in the cloud manufacturing platform.

A proposed SLA framework for service sharing in virtual environments is given; electronic contracts are designed in the framework for encapsulating measurable aspects of service level agreements so as to provide common understanding about the service; and an improved SLA negotiation strategy with three phases is presented for the dynamicity of the virtual services.

The paper presents a very useful adaptive SLA mechanism for service sharing in virtual environments that can be utilized in concurrent or future advanced manufacturing modes.