检索结果-南通市图书馆

Automated classification and detection of multiple pavement distress images based on deep learning

在线全文

学校读者我要写书评

暂无评论

Journal of Traffic and Transportation Engineering(English Edition) 2023年第2期10卷 276-290页

作者： Deru Li Zhongdong Duan xiaoyang hu Dongchang Zhang Yiying Zhang School of Civil and Environmental Engineering Harbin Institute of Technology(Shenzhen)Shenzhen 518055China China Merchants Roadway Information Technology(Chongqing)Co.Ltd. Chongqing 400067China

To achieve automatic,fast,efficient and high-precision pavement distress classification and detection,road surface distress image classification and detection models based on deep learning are ***,a pavement distress image dataset is built,including 9017pictures with distress,and 9620 pictures without *** pictures were captured from 4 asphalt highways of 3 provinces in *** each pavement distress image,there exists one or more types of distress,including alligator crack,longitudinal crack,block crack,transverse crack,pothole and *** distresses are labeled by a rectangle bounding box on the *** ResNet networks and VGG networks are used respectively as binary classification models for distressed and non-distressed imagines classification,and as multi-label classification models for six types of distress *** techniques,such as data augmentation,batch normalization,dropout,momentum,weight decay,transfer learning,and discriminative learning rate are used in training the *** the 4 CNNs considered in this study,namely ResNet 34 and 50,and VGG 16 and 19,for the binary classification,ResNet 50 has the highest Accuracy of 96.243%,Precision of 95.183%,and ResNet 34 has the highest Recall of 97.824%,and F2 score of 97.052%.For multi-label classification,ResNet 50 has the best performance,with the highest Accuracy of 90.257%,higher than 90%required by the Chinese standard(JTG H20-2018)for road distresses detection,F2 score-82.231%,and Precision-76.509%,and ResNet34 has the highest Recall of 87.32%.To locate and quantify the distress areas in the images,the single shot multibox detector(SSD)model is developed,in which the ResNet 50 is used as the base network to extract *** the intersection over union(IoU)is set to 0,0.25,0.50,0.75,the mean average precision(mAP)of the model are found to be 74.881%,50.511%,28.432%,3.969%,respectively.

关键词： Pavement distress Deep learning Multi-label classification Distress detection Single shot multibox detector Convolutional neural network

Preparation and properties of dental zirconia ceramics

在线全文

学校读者我要写书评

暂无评论

Journal of University of Science and Technology Beijing 2008年第6期15卷 764-768页

作者： Xinjie Liang Yuexiu Qiu Shaoxiong Zhou xiaoyang hu Guangyan Yu Xuliang Deng China Iron and Steel Research Institute Group Beijing 100081 China School and Hospital of Stornatology Peking University Beijing 100081 China

Y2O3-stabilized tetragonal zirconia polycrystalline （Y-TZP） ceramics with high-performance were prepared for dental ap- plication by use of the micro-emulsion and two-step sintering method. The crystal phase, morphology, and microstructure of the reaction products were characterized by X-ray diffraction （XRD）, scanning electron microscopy （SEM）, and transmission electron mi- croscopy （TEM）. XRD results show that the ceramics mainly consist of tetragonal zirconia. Physical and mechanical properties test results show that the bending strength, fracture toughness, and the density of full sintered Y-TZP ceramics are 1150 MPa, 5.53 MPa.m1/2, and 6.08 g/cm3, respectively, which suggest that the material is relatively suitable for dental restoration. The dental base crown machined with this material by CAD/CAM system exhibits a verisimilitude configuration and the material＇s expansion coefficient well matches that of the glaze. These results further indicate that the product can be used as a promising new ceramic material to fabricate dental base crowns and bridges.

关键词： zirconia ceramic crown property

Automated classification and detection of multiply pavement distress images based on Deep Learning

在线全文

学校读者我要写书评

暂无评论

Journal of Traffic and Transportation Engineering（English Edition） 2021年

作者： Deru Li Zhongdong Duan xiaoyang hu Dongchang Zhang Yiying Zhang School of Civil and Environmental Engineering， Harbin Institute of Technology China Merchants Roadway Information Technology (Chongqing) CO.LTD

To achieve automatic， fast， efficient and high-precision pavement distress classification and detection， road surface distress image classification and detection models based on deep learning are trained. First， a pavement distress image dataset is built， including 9017 pictures with distress， and 9620 pictures without distress. These pictures were captured from 4 asphalt highways of 3 provinces in China. In each pavement distress image， there exists one or more types of distress， including alligator crack， longitudinal crack， block crack， transverse crack， pothole and patch. The distresses are labeled by a rectangle bounding box on the pictures. Then ResNet networks and VGG networks are used respectively as binary classification models for distressed and non-distressed imagines classification， and as multi-label classification models for six types of distress classification. Training techniques， such as data augmentation， batch normalization， dropout， momentum， weight decay， transfer learning， and discriminative learning rate are used in training the model. Among the 4 CNNs considered in this study， namely ResNet 34 and 50， and VGG 16 and 19， for the binary classification， ResNet 50 has the highest Accuracy of 96.243%， Precision of 95.183%， and ResNet 34 has the highest Recall of 97.824%， and F2 score of 97.052%. For multi-label classification， ResNet 50 has the best performance， with the highest Accuracy of 90.257%， higher than 90% required by the Chinese standard （JTG H20-2018） for road distresses detection， F2 score - 82.231%， and Precision - 76.509%， and ResNet 34 has the highest Recall of 87.32%.To locate and quantify the distress areas in the images， an SSD （Single Shot MultiBox Detector） model is developed， in which the ResNet 50 is used as the base network to extract features. When the IoU （Intersection over Union） is set to 0.0， 0.25， 0.5， 0.75， the mean Average Precision （mAP） of the model are found to be 74.881%， 50.511%， 28.432%， 3.969%， respectively.

关键词： Pavement Distress Deep Learning Multi-label Classification Distress Detection Single Shot MultiBox Detector Convolutional Neural Network

同方期刊数据库

Key elements required to publish a good article:consensus among executive editors of Traditional Chinese Medicine and integrative medicine journals——a Delphi study

在线全文

学校读者我要写书评

暂无评论

Journal of Traditional Chinese Medicine 2015年第3期35卷 335-342页

作者： hu xiaoyang Robinson Nicola Yu He Liu Jianping School of Health and Social Care London South Bank University School of Basic Medicine Beijing University of Chinese Medicine Centre for Evidence-Based Chinese Medicine Beijing University of Chinese Medicine

OBJECTIVE:To explore consensus among Traditional Chinese Medicine(TCM)/ integrative medicine(IM) journal executive editors in China regarding:(a)Their perceptions about the differences between papers published in Chinese/English journals;(b)the key elements required for publishing a good quality paper;(c) how editors manage and decide on the manuscripts submitted to their ***:A three classic round Delphi was conducted with a purposive sample of TCIWIM Chinese journal executive *** key themes and items generated from six open ended questions in the first round were redistributed to the editors in two subsequent questionnaire rounds and scored on a five-point Likert ***,reliability and stability of responses were assessed between the editors in the same round and between ***:All ten editors initially agreeing to participate completed the three rounds,provided 60 comments in the first round,relating to 6 themes on:elements of publishing good quality article;differences between Chinese and English published articles;reasons for low quality of Chinese studies;maintaining the journal's good reputation;initiatives and strategies;and differences between Chinese and English versions of the same *** levels of consensus were reached for 97%***-class correlation(ICC) coefficient for agreement among editors was high in both second and third round and there was a good stability between the two ***:High inter/intra-editor agreement was identified in the differences between TCM/IM research papers published in English and *** standards and probity are keys in elevating the reputation of ***' perspectives in maintaining and improving quality journal can be viewed as fundamental for editors and researchers in publishing any articles.

关键词： Medicine,Chinese Traditional Integrative medicine Delphi Publication quality Editors' perspectives

Recent Progress in Fiber-Optic Hydrophones

在线全文

学校读者我要写书评

暂无评论

Photonic Sensors 2021年第1期11卷 109-122页

作者： Zhou MENG Wei CHEN Jianfei WANG xiaoyang hu Mo CHEN Yichi ZHANG College of Meteorology and Oceanology National University of Defense TechnologyChangsha410073China Academy of Artillery and Air Defense Nanjing210000China

Fiber-optic hydrophone (FOH) is a significant type of acoustic sensor, which can be used in both military and civilian fields such as underwater target detection, oil and natural gas prospecting, and earthquake inspection. The recent progress of FOH is introduced from five aspects, including large-scale FOH array, very-low-frequency detection, fiber-optic vector hydrophone (FOVH), towed linear array, and deep-sea and long-haul transmission. The above five aspects indicate the future development trends in the FOH research field, and they also provide a guideline for the practical applications of FOH as well as its array.

关键词： Fiber-optic hydrophone large-scale array very-low-frequency detection fiber-optic vector hydrophone towed linear array deep sea long-haul fiber transmission

Evidence base of randomized controlled trials and guideline recommendations of patent Traditional Chinese Medicines for uncomplicated acute lower respiratory tract infections in adults

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Traditional Chinese Medicine 2018年第4期38卷 490-503页

作者： Xia Ruyu hu xiaoyang Wang Di Wang Ying Merlin Willcox Liu Jianping Michael Moore Andrew Flower Li Xun Lily Lai hu Ruixue Wen Lingzi Zhang Lishan Wang Qi Fei Yutong Centre for Evidence-Based Chinese Medicine Beijing University of Chinese Medicine Primary Care and Population Sciences University of Southampton Respiratory Department Dongzhimen Hospital Affiliated to Beijing University of Chinese Medicine Respiratory Department Dongfang Hospital Affiliated to Beijing University of Chinese Medicine

OBJECTIVE:To compare clinical practice guideline recommendations on the use of oral patent Traditional Chinese Medicines(PTCMs) for uncomplicated acute lower respiratory tract infections(ALRTIs)in adults with the existing evidence using results of a systematic review of randomized controlled trials(RCTs).METHODS:A systematic review on RCTs and a systematic review of current guidelines on orally taken PTCMs for uncomplicated ALRTIs were *** Med,Cochrane Library,EMBASE and four Chinese databases were searched from inception to September 2016 for RCTs testing orally taken PTCMs for uncomplicated ALRTIs(excluding pneumonia).Two reviewers independently screened each study,extracted study data,and assessed risk of *** were resolved through discussion or by consultation with a third *** practice guidelines for uncomplicated ALRTIs containing PTCM recommendations were identified and quality *** quality of pooled evidence of the RCTs and the guidelines was assessed with GRADE and AGREE Ⅱ *** consistency of the evidence base in RCTs and the guideline recommendations were then ***:For the systematic review of RCTs,4810 papers were identified,among which 29 RCTs(5093 patients) were included in the *** compared to placebo increased the effective treatment rate of cough(3 trials,949 patients,risk ratio(RR) 2.50,1.16 to 5.43;low certainty);improved assessment of global health(3 trials,948 patients,RR1.70,1.44 to 2.01;low certainty);and increased the effective rate of specific symptom relief(1 trial,478 patients,RR 4.01,2.76 to 5.81;moderate certainty).21 trials(3432 patients) compared effects of different *** the guideline evaluation,29 PTCMs were recommended for the use of uncomplicated ALRTIs,of which27 had no supportive evidence from ***:The evidence base of PTCMs for uncomplicated ALRTIs is weak and the guideline recommendations were based on almost no clinical trial evidence.

关键词： Traditional Chinese Medicine Acutelower respiratory tract infections Antibiotic Ay-stematic review Guideline

Synthesis of Fe3O4 Nanoparticles and their Magnetic Properties

在线全文

学校读者我要写书评

暂无评论

Synthesis of Fe3O4 Nanoparticles and their Magnetic Properti...

2011中国材料研讨会

作者： Yan Wei,Bing Han,xiaoyang hu

＜正＞Fe3O4 magnetic nanoparticles （MNPs） were synthesized by a co-precipitation method using odiumcitrate and oleic acid as *** composition and microstructure analysis indicate that the sodium citrate

关键词：

来源： cnki会议评论

在线全文

cnki会议

学校读者我要写书评

暂无评论

我国高校排舞课程发展困境与建设路径研究

沈阳体育学院学报 2018年第5期37卷 125-130页

作者：胡晓洋四川外国语大学体育部重庆410031

运用文献资料、实地调研、逻辑分析等方法,在高校排舞课程建设学理分析的基础上,对我国高校排舞课程建设成效、发展困境以及建设路径进行研究。研究结果表明:当前我国高校排舞面临着课程普及程度不高、教师专业水平偏低、部分学生学习... 详细信息

运用文献资料、实地调研、逻辑分析等方法,在高校排舞课程建设学理分析的基础上,对我国高校排舞课程建设成效、发展困境以及建设路径进行研究。研究结果表明:当前我国高校排舞面临着课程普及程度不高、教师专业水平偏低、部分学生学习效果欠佳等困境,为切实有效破解困境,提出高校排舞课程建设具体的路径:倡导"健康第一"思想形塑课程实施主体理念,建立同业交流平台扩大排舞课程普及,优化教师配置提升排舞师资水平,完善评价机制提高学生排舞学习效果,结合排舞特点加强课程整体设置。

关键词：高校排舞课程 “健康第一”思想课程建设课程普及课程整体设置

聚四氟乙烯微粉对改性氟橡胶摩擦磨损性能的影响

在线全文

学校读者我要写书评

暂无评论

高分子材料科学与工程 2019年第7期35卷 69-73,80页

作者：胡晓阳帅长庚杨雪海军工程大学振动与噪声研究所湖北武汉430033 船舶振动噪声重点实验室湖北武汉430033

通过加入聚四氟乙烯(PTFE)微粉对氟橡胶进行改性,采用力学试验和环-块摩擦磨损试验研究了不同含量的聚四氟乙烯对氟橡胶的力学性能和摩擦磨损性能的影响,使用扫描电子显微镜分析了磨损机制,并结合力学性能讨论了摩擦磨损性能变化的原因... 详细信息

通过加入聚四氟乙烯(PTFE)微粉对氟橡胶进行改性,采用力学试验和环-块摩擦磨损试验研究了不同含量的聚四氟乙烯对氟橡胶的力学性能和摩擦磨损性能的影响,使用扫描电子显微镜分析了磨损机制,并结合力学性能讨论了摩擦磨损性能变化的原因。结果表明,随着PTFE含量增加,硫化胶的拉伸强度降低但撕裂强度升高;PTFE质量含量在8%以下的改性氟橡胶摩擦磨损性能得以改善,体积磨损率和摩擦系数均小于未改性氟橡胶;5%的PTFE改性氟橡胶耐磨性最好,10%的PTFE改性氟橡胶摩擦系数最小。

关键词：聚四氟乙烯氟橡胶摩擦磨损