检索结果-内蒙古大学图书馆

Distilling object detectors with mask-guided feature and relation-based knowledge

INTERNATIONAL JOURNAL OF COMPUTATIONAL SCIENCE AND engineering 2024年第2期27卷 195-203页

作者： Zeng, Liang Ma, Liyan Luo, Xiangfeng Guo, Yinsai Chen, Xue Shanghai Univ Sch Comp Engn & Sci Shanghai 200444 Peoples R China State Key Lab Math Engn & Adv Comp Wuxi 214083 Peoples R China

Knowledge distillation (KD) is an effective technique for network compression and model accuracy enhancement in image classification, semantic segmentation, pre-trained language model, and so on. However, existing KD methods are specialised for image classification and cannot be used effectively for object detection tasks, with the following two limitations: the imbalance of foreground and background instances and the neglect distillation of relation-based knowledge. In this paper, we present a general mask-guided feature and relation-based knowledge distillation framework (MAR) consisting of two components, mask-guided distillation, and relation-based distillation, to address the above problems. The mask-guided distillation is designed to emphasise students' learning of close-to-object features via multi-value masks, while relation-based distillation is proposed to mimic the relational information between different feature pixels on the classification head. Extensive experiments show that our methods achieve excellent AP improvements on both one-stage and two-stage detectors. Specifically, faster R-CNN with ResNet50 backbone achieves 40.6% in mAP under 1 x schedule on the COCO dataset, which is 3.2% higher than the baseline and even surpasses the teacher detector.

关键词： knowledge distillation multi-value mask object detection

来源：评论

学校读者我要写书评

暂无评论

Robust Solid-state Na-CO2 Battery with Na2.7Zr2Si2PO11.7F0.3-PVDF-HFP Composite Solid Electrolyte and Na15Sn4/Na Anode

引用

ACS APPLIED MATERIALS & INTERFACES 2024年第10期16卷 12706-12716页

作者： Wang, Zelin Mao, Yuezhen Sheng, Lunhuai Sun, Chunwen China Univ Min & Technol Beijing Sch Chem & Environm Engn Beijing 100083 Peoples R China

Solid-state Na-CO2 batteries are a kind of energy storage devices that can immobilize and convert CO2. They have the advantages of both solid-state batteries and metal-air batteries. High-performance solid electrolyte and electrode materials are important for improving the performance of solid-state Na-CO2 batteries. In this work, we investigate the influence of fluorine doping on the structure and ionic conductivity of Na3Zr2Si2PO12 (NZSP). An ionic conductive solid electrolyte membrane was prepared by compositing the inorganic solid electrolyte Na2.7Zr2Si2PO11.7F0.3 (NZSPF3) with poly(vinylidene fluoride)-co-hexafluoropropylene (PVDF-HFP). It shows an ionic conductivity of up to 2.17 x 10(-4) S cm(-1) at room temperature, a high sodium ionic transfer number of similar to 0.70, a broad electrochemical window of similar to 5.18 V, and better mechanical strength. Furthermore, we studied the Na15Sn4/Na composite foil with the ability to inhibit dendrite as the anode for solid-state Na-CO2 batteries. Through density functional theory (DFT) calculations, the Na15Sn4 particle has been verified with a strong sodiophilic property, which reduces the nucleation barrier during the deposition process, leading to a lower overpotential. The symmetric cell assembled with the composite solid-state electrolyte NZSPF3-PVDF-HFP and Na15Sn4/Na composite anode can inhibit the growth of Na dendrites effectively and maintain the stability of the whole cell structure. Solid-state Na-CO2 batteries assembled with Ru-carbon nanotube (Ru-CNTs) as cathode catalysts exhibit a high discharge capacity of 6371.8 mAh g(-1) at 200 mA g(-1), excellent cycling stability for 1100 h, and good rate performance. This work provides a promising strategy for designing high-performance solid-state Na-CO2 batteries.

关键词： solid-state Na-CO2 batteries composite solidelectrolyte NASICON F doping Na-Snalloy anode

来源：评论

学校读者我要写书评

暂无评论

Study on the Wear Resistance of TiN/TiO2 Micro-arc Oxidized Composite Coating on TC4 Titanium Alloy

引用

JOURNAL OF MATERIALS engineering AND PERFORMANCE 2025年 1-12页

作者： Chen, Xiaowen Xie, Wanlin Wang, Ling Tang, Song Zhang, Defen Southwest Petr Univ Sch New Energy & Mat Chengdu 610500 Sichuan Peoples R China

To improve the abrasion resistance of TC4 (Ti-6Al-4V) titanium alloy components, TiN/TiO2 composite ceramic coating was prepared on the surface of TC4 titanium alloy by introducing titanium nitride (TiN) particles into the silicate-phosphoric acid composite electrolyte system using the micro-arc oxidation (MAO) technique. The surface morphology, phase composition, and roughness of the coating were analyzed by SEM, an x-ray diffractometer, and a handheld roughness instrument, respectively. The abrasion resistance of the coating prepared by the TiN additive was analyzed by a friction abrasion tester. These results demonstrate that the coatings formed in the electrolyte consist primarily of rutile-TiO2 phase and anatase-TiO2 phase with a dense structure. During the MAO process, TiN successfully penetrates the coating and reduces the average size and porosity of the pores by filling the micropores on the coating surface by mechanical agitation and electrophoresis. The MAO coatings achieved their highest thickness and hardness when 3 g/L TiN particles were added, as well as a 43% reduction in the coefficient of friction when compared to the titanium alloy substrate and a relatively lower abrasion width of the coating that has good wear.

关键词： friction wear micro-arc oxidation TC4 titanium alloy TiN additions

来源：评论

学校读者我要写书评

暂无评论

A two-stage multi-attribute analysis method for city-integrated hybrid mini-grid design

引用

SUSTAINABLE CITIES AND SOCIETY 2021年 65卷

作者： Elkadeem, M. R. Kotb, Kotb M. Ullah, Zia Atiya, Eman G. Dan, Andras Wang, Shaorong Huazhong Univ Sci & Technol Sch Elect & Elect Engn Wuhan 430074 Peoples R China Budapest Univ Technol & Econ Fac Engn & Informat Dept Elect Power Engn H-1111 Budapest Hungary Tanta Univ Elect Power & Machines Engn Dept Fac Engn Tanta 31521 Egypt

This paper proposes a two-stage multi-attribute analysis method in conceptualizing the feasible electrification strategy with optimal capacities for a mid-rise building situated in El-Qsier urban city, Egypt. The proposed model measures the implementation practicability of hybrid mini-grid against stand-alone diesel and the utility grid extension systems. The model takes into consideration, simultaneously, a total of 12 attributes covering technical, economic, environmental, and socio-political sustainability aspects. First, an accurate Energy-EnviroEconomic (3E) optimization analysis is performed to determine the set of feasible configurations using HOMER Pro software. Second, a Multi-Attribute Decision Making (MADM) model is established based on AHP, TOPSIS, VIKOR, CODAS, WASPAS methods to identify the unique best alternative. The results of the 3E optimization and decision making analyses confirm the non-viability of the reference site for grid extension against other off-grid systems. Also, the 100 % hybrid renewable mini-grid system is nominated as the most sustainable design, and it yielded optimal capacities of 196 kW, 72 kW, 587 kW h, and 75 kW for photovoltaic, wind turbine, converter, and battery. The system has total life-cycle and energy costs of 751,597$ and 0.2374$/kWh, respectively, with a zero-emission value and maximum social benefits equals 0.6089 jobs/year. Meanwhile, the sensitivity analysis elucidates that the load growth and battery cost have a high impact on investment decisions rather than solar irradiance and wind speed.

关键词： Hybrid mini-grid energy system Energy-enviro-economic optimization Grid extension Multi-attribute decision making analysis Sensitivity analysis

来源：评论

学校读者我要写书评

暂无评论

FeCo 5 /Nitrogen doped carbon as an efficient bifunctional oxygen electrocatalyst for Zn-Air batteries

引用

JOURNAL OF ELECTROANALYTICAL CHEMISTRY 2024年 965卷

作者： Wang, Jingyu Zhang, Tianai He, Shengzhi Sun, Chunwen China Univ Min & Technol Beijing Sch Chem & Environm Engn Beijing 100083 Peoples R China

Bifunctional non-precious metal catalysts (NPMCs) are crucial for accelerating the oxygen reduction reaction (ORR) and oxygen evolution reaction (OER) at cathode of zinc -air batteries (ZABs). In this work, we report bimetallic FeCo x /nitrogen doped carbon (NC) catalysts synthesized through a metal-organic frameworks (MOFs) confinement strategy to ensure uniform dispersion of Fe and Co, followed by high-temperature pyrolysis. The prepared FeCo x /nitrogen doped carbon catalyst is rich in Fe-N x and Co-N x active sites and exhibits excellent electrocatalytic activities towards ORR and OER in alkaline electrolyte. Among compositions studied, FeCo x / nitrogen doped carbon catalyst demonstrates a high half-wave potential ( E 1/2 ) of 0.908 V, an overpotential of 440 mV at E j=10 , and a low potential difference ( Delta E ) between E j=10 and E 1/2 of only 0.762 V. After a 10 h chronoamperometry (CA) test at an applied voltage of 0.5 V, the catalyst retains 91 % of its initial activity. The synthesized catalyst also shows practical utility when used as cathodic catalysts in ZABs, exhibiting a high peak power density of 155.7 mW cm -2 , a excellent discharge specific capacity of 715.1 mAh g -1 and long cycle stability for 188 h.

关键词： Non-precious Metal Catalyst Zinc-Air Battery Oxygen Reduction Reaction Oxygen Evolution Reaction Bimetallic Catalyst

来源：评论

学校读者我要写书评

暂无评论

Effect of Mesophase Pitch Incorporation on the Ablation Behavior and Mechanism of Phenolic Composites

引用

INDUSTRIAL & engineering CHEMISTRY RESEARCH 2022年第13期61卷 4612-4624页

作者： Huang, Yisen Zhang, Haoruo Zhang, Xueqin Yan, Liwei Ling, Youquan Zou, Huawei Chen, Yang Liang, Mei Sichuan Univ Polymer Res Inst State Key Lab Polymer Mat Engn Chengdu 610065 Peoples R China

Phenolic resin plays a critical role in ablative thermal protection materials. However, its ablation and insulation properties need to be further improved to meet the growing requirements of the entire aerospace industry. In this work, composites are altered with various loadings of mesophase pitch (MP) to make mesophase pitch-modified boron-phenolic resin (BPR) composites (MPBPR). The compressive strengths of carbonized MPBPR composites are significantly improved (optimally by 2.7 times) via this modification, indicating the improved mechanical performance of MPBPR composites under high temperatures. Results reveal that the obtained MPBPR composites show better ablation resistance and thermal insulation performance under oxyacetylene flame. The linear ablation rates (4 MW/m2, lasting for 30 s) of the composites with 10 and 20 wt % MP are reduced by 41.5 and 55.4% compared to that of neat BPR. Meantime, MPBPR composites exhibit lower back-face temperature during the ablation process. Furthermore, results also show that the introduction of MP can effectively avoid the growth of large transversal microcracks, increase the proportion of micropores in the ablated char layer, and simultaneously improve the graphitization of resultant char. Both the microporous and graphitic structures of the residual char contribute to the improved ablation and insulation properties of MPBPR composites. The abovementioned results indicate that these MP-modified boron-phenolic resins are promising materials for application in a thermal protection system.

关键词： Hydrocarbons Ablation Composites Layers Aromatic compounds

来源：评论

学校读者我要写书评

暂无评论

Few-Shot Hyperspectral Image Classification Using Relational Generative Adversarial Network

引用

IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING 2024年 62卷

作者： Wei, Wei Tong, Lei Guo, Baoqing Zhou, Jun Xiao, Chuangbai Beijing Univ Technol Coll Comp Sci Beijing 100124 Peoples R China Beijing Jiaotong Univ State Key Lab Adv Rail Autonomous Operat Beijing 100044 Peoples R China Griffith Univ Sch Informat & Commun Technol Nathan Qld 4111 Australia

Hyperspectral image (HSI) classification is an essential task in remote sensing, but its performance is greatly affected by limited labeled samples. Currently, generative adversarial network (GAN)-based methods can generate the virtue samples to augment the training set. However, with limited labeled data, GANs perform poorly in capturing features during sample generation. Very few relation networks (RNs) and few-shot learning (FSL) methods considered data augmentation to enhance performance. To address this challenge, we propose FSHyperRGAN, a few-shot HSI classification method based on relational GAN, which uses GANs to augment the training samples for RNs, while leveraging relation feature extraction to guide the generation of specific class samples. FSHyperRGAN comprises four modules: a data processing module converting the HSI data to 1-D and 3-D features, an adversarial generation (AG) module synthesizing virtual samples conditioned on labels, a data embedding and reconstruction (DER) module encoding latent spaces for accurate sample reconstruction while preserving category characteristics, and a relation computation (RC) module computing relation scores across generated, reconstructed, and original samples. In addition, a relational feature matching scheme is also applied, which can use virtual samples to guide classification. Two FSHyperRGAN frameworks are designed, 1D-FSHyperRGAN and 3D-FSHyperRGAN, which can be utilized for 1-D spectral or 3-D spatial-spectral classification. Experiments on widely used HSI datasets illustrate that the proposed method outperforms several state-of-the-art methods and achieves overall accuracies of 92.64%, 86.82%, 83.64%, and 84.57% on KSC, PaviaU, Houston, and WHU-Hi-HongHu datasets, respectively.

关键词： Feature extraction Generative adversarial networks Few shot learning Training Accuracy Image reconstruction Hyperspectral imaging Three-dimensional displays Generators Image classification Deep learning few-shot learning (FSL) generative adversarial network (GAN) hyperspectral image (HSI) classification

来源：评论

学校读者我要写书评

暂无评论

A buried interface modification strategy for enhancing the photovoltaic performance of NiOx-based inverted perovskite solar cells

引用

VACUUM 2024年 222卷

作者： Wu, Yue Wei, Mengyuan Sun, Yuxuan Yang, Xiao Xun, Wei Li, Mingguang Chen, Runfeng Lin, Yi Li, Ping Jiang, Qing-Song Huaiyin Inst Technol Fac Elect Informat Engn Huaian 223003 Peoples R China Nanjing Univ Posts & Telecommun State Key Lab Organ Elect & Informat Displays Nanjing 210023 Peoples R China Nanjing Univ Posts & Telecommun Inst Adv Mat Nanjing 210023 Peoples R China Huaiyin Inst Technol Fac Math & Phys Huaian 223003 Peoples R China Xi An Jiao Tong Univ Sch Mat Sci & Engn Xian 710049 Shaanxi Peoples R China

Nickel oxide (NiOx) is considered to be an attractive hole transport material in planar inverted perovskite solar cells (PSCs). Unfortunately, the further development of NiOx-based PSCs is severely restrained due to the poor interface quality between NiOx and perovskite layers. In this paper, a buried interface modification strategy using a polar molecule material 4,4-difluorocyclohexylamine hydrochloride (simplified as DFCHY) is introduced to improve the performance of NiOx-based PSCs. Theoretical analysis and experimental results demonstrate that DFCHY plays a role in bridging the NiOx/perovskite interface, not only suppressing the high oxidation state of Ni3+ and enhancing the conductivity in NiOx film, but also passivating buried interface of perovskite film and improving its crystallinity, further effectively reducing trap density and enhancing charge transfer at the interface of NiOx/perovskite layers. The champion inverted PSCs using DFCHY as a buried interface modifier achieve a power conversion efficiency of 17.08% with negligible hysteresis. Furthermore, the unencapsulated PSCs with DFCHY modification exhibit outstanding long-term stability, maintaining above 90% of their initial efficiency after 1000 h storage in a nitrogen-filled glovebox.

关键词： Nickel oxide

来源：评论

学校读者我要写书评

暂无评论

UWB-IMU-Odometer Fusion for Simultaneous Calibration and Localization

引用

IEEE INTERNET OF THINGS JOURNAL 2025年第1期12卷 950-963页

作者： Sun, Jian Sun, Wei Zheng, Jin Fang, Xu Liu, Jian Mian, Ajmal Hunan Univ Coll Elect & Informat Engn Natl Engn Res Ctr Robot Visual Percept & Control T Changsha 410082 Peoples R China Hunan Univ Coll Elect & Informat Engn Changsha 410082 Peoples R China Hunan Univ Sch Robot Changsha 410082 Peoples R China Cent South Univ Sch Architecture & Art Changsha 410082 Peoples R China Nanyang Technol Univ Sch Elect & Elect Engn Singapore Singapore Univ Western Australia Dept Comp Sci Crawley WA Australia

The location accuracy of fixed anchors plays a pivotal role in ultrawideband (UWB) positioning. However, existing calibration methods for calculating the anchor positions require anchor-to-anchor communication or known initial values for anchor locations. Moreover, the calibration accuracy is adversely affected by non line-of-sight (NLOS) conditions. We propose a two-stage calibration scheme to conduct simultaneous calibration and localization (SCAL) based on UWB-inertial measurement unit (IMU)-odometer sensor fusion without the need for anchor-to-anchor communication or manual intervention. Our method performs IMU-odometer-aided multidimensional scaling (IO-MDS) to provide the initial calibration value without anchor-to-anchor ranging measurements. This is followed by a novel factor-graph-based framework to achieve coarse-to-fine calibration based on UWB, inertial, and odometer measurements. In existing works, a single UWB range measurement is regarded as a weak constraint, as it often leads to incorrect estimates. Our method uses the derived radial velocity (DRV) and IO-MDS factors as additional strong constraints for reliable estimates. To minimize the NLOS influence on the calibration process, we introduce an improved least square-support vector machine (ILS-SVM) based on adaptive weight parameter and multikernel function. Experimental results on two field collected datasets show enhancements in NLOS identification and SCAL. In the lab dataset, identification accuracy increased by 5.5%, with improvements of 0.124 m and 0.309 m in root mean square error (RMSE) for robot and anchor locations, respectively. In the parking lot dataset, identification accuracy improved by 5.0%, with RMSE improvements of 0.223 m and 0.317 m for robot and anchor locations, respectively.

关键词： Calibration Accuracy Location awareness Position measurement Nonlinear optics Sun Odometers Wireless sensor networks Manuals Internet of Things factor graph localization non line-of-sight (NLOS) ultrawideband (UWB)/inertial measurement unit (IMU)/odometer

来源：评论

学校读者我要写书评

暂无评论

Effect of Heat Treatment on Tribological Properties of SiCp/Al-Mg-Mn-Cu Aluminum Matrix Composites

引用

TRANSACTIONS OF THE INDIAN INSTITUTE OF METALS 2024年第9期77卷 2679-2686页

作者： Zhou, Xie Li, Fangjie Liu, Min Yang, Dongye Shen, Qin Zheng, Jun Xiong, Changhong Zhang, Yan Shanghai Univ Engn Sci Sch Mat Sci & Engn 333 Longteng Rd Shanghai 201602 Peoples R China Shanghai Inst Technol Sch Mat Sci & Engn 100 Haiquan Rd Shanghai 201418 Peoples R China Shanghai Univ Engn Sci Sch Mech & Automot Engn 333 Longteng Rd Shanghai 201602 Peoples R China Anhui Univ Technol Key Lab Green Fabricat & Surface Technol Adv Met Minist Educ Maanshan 243002 Peoples R China

The tribological behaviors of SiCp/Al-Mg-Mn-Cu aluminum matrix composites under solid solution (labeled as SS sample) and aging treatment (labeled as SA sample) were systematically investigated. The microstructures, tribological properties and wear mechanisms of worn surfaces were analyzed using X-ray diffraction (XRD), scanning electron microscopy (SEM) and wear testing machine, respectively. The microscopy study showed the uniform distribution of reinforcements in the SiCp/Al-Mg-Mn-Cu composites. And a large number of fine CuMgAl2 phases were observed in the untreated sample but dissolved in the SS sample to some extent. It can be confirmed that the size and number of some new precipitates and CuMgAl2 phases were in the suitable state for the SA sample resulting from its excellent integrated mechanical properties. A good combination of strength and plasticity in SA sample was corresponding well with its much lower coefficient of friction (COF) and well abrasion resistance, with a COF of 0.25. The wear mechanisms of the untreated and SS samples have found to be the combination of abrasive and adhesive indicated by the presence of some wear debris and delaminated flakes, whereas the wear mechanism is changed to adhesive dominantly with slight delamination in case of SA sample.

关键词： SiCp/Al-Mg-Mn-Cu composites Microstructure Wear resistance Wear mechanism

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：