Results 51 to 60 of about 815 (137)

Review of Deep Learning Applications in Unmanned Aerial Vehicle Remote Sensing Images of Grass Plants [PDF]

open access: yes
The invasion of grass species competes for resources necessary for the growth of crops, significantly affecting crop yield and quality. Grasses such as the Artemisia genus release allergenic pollen, causing allergic reactions and impacting human health ...
LI Shaobo, WANG Xiaoqiang, GUO Libiao, HONG Ying, WANG Zhiguo
core   +1 more source

电力物联网下基于卷积神经网络和迁移学习的GIS局部放电模式识别分类方法研究

open access: yesGaoya dianqi, 2020
随着电力物联网建设的快速推进,对实时监测气体绝缘金属封闭开关设备(GIS)内部局部放电信号的特高频法提出了新的和更高的要求。在充分利用表征GIS局部放电信息构建模型以提升模式识别准确率的同时,将模型移植到嵌入式系统,进而构成以边缘计算为支撑的物联网智能终端,成为亟待解决的一个关键问题。为此,文中深入研究了基于卷积神经网络和迁移学习的GIS局部放电模式识别分类方法,在多种典型缺陷下对比了不同模型在训练时间、准确率、参数量和存储花销等方面的性能。结果表明 ...
杨为   +8 more
doaj  

Timing data visualization: tactical intent recognition and portable framework [PDF]

open access: yes
By transforming time series into images, a robust and transferable tactical intent recognition framework was proposed, which integrated curve filtering technology and the EfficientNetV2 image recognition network.
LI Lemin   +4 more
core   +1 more source

白鹇海南亚种的行为时间分配及其活动节律

open access: yes野生动物学报, 2021
为掌握白鹇海南亚种(Lophura nycthemera whiteheadi)的生态习性,探讨有效保护策略,2018年1月—2021年1月,在海南热带雨林国家公园五指山片区内布设50台红外相机进行连续监测,共收集白鹇独立有效照片1 472张,视频491个,对其行为时间分配,活动节律(日、月、季节)与海拔迁移规律进行分析。结果表明:①白鹇的日活动时间从06:00开始到20:00结束,其中出现3个活动高峰,即07:00—09:00、13:00和16:00;雄性与雌性的日活动频率无显著差异(t=2020 ...
李佳灵 1   +1 more
doaj  

Computer Vision-assisted Wireless Channel Simulation for Millimeter Wave Human Motion Recognition [PDF]

open access: yes
This study proposes a computer vision-assisted millimeter wave wireless channel simulation method incorporating the scattering characteristics of human motions.
Chao YU   +4 more
core   +1 more source

基于循环微调训练的柴油机故障迁移学习诊断

open access: yesZhendong Ceshi yu Zhenduan
针对当前柴油机故障诊断领域深度学习模型参数规模大、训练时间长,以及工程应用中带标签样本不足的问题,提出一种基于轻量化网络和循环微调训练的柴油机故障诊断迁移学习方法。首先,该方法采用轻量化MobileNet‑V2作为主干网络,在权重迁移学习的基础上,提出一种部分权重循环初始化微调的方法,解决了变工况条件下的小样本诊断问题;其次,开展不同程度进气滤清器堵塞和气门间隙异常的柴油机故障台架实验;最后,采用所提方法和现有方法,对变负荷条件下的柴油机进行故障诊断对比分析。结果表明:所提方法在参数规模、计算量 ...
doaj   +1 more source

丹戎布拉国立大学师范教育学院汉语专业2015届学生 使用名词“后来”和“以后”的偏误分析 [PDF]

open access: yes, 2017
In Mandarin learning, grammar has a very important role. The noun is one of the most important parts of grammar, including the noun "houlai" and "yihou". "Houlai" and "yihou" also include the word synonym. Synonyms are words that contain the same meaning.
Mayasari, A. R. (A‘an)
core  

Automated Static Warning Identification via Path-based Semantic Representation

open access: yes, 2023
Despite their ability to aid developers in detecting potential defects early in the software development life cycle, static analysis tools often suffer from precision issues (i.e., high false positive rates of reported alarms).
Jin, Zhi   +3 more
core  

基于改进YOLOv4的配电线路绝缘子与避雷器快速检测研究

open access: yesDianci bileiqi, 2023
配电线路中绝缘子与避雷器的故障影响供电系统的可靠性,为了实现航拍图像中绝缘子与避雷器目标的快速、精准检测,笔者提出一种基于改进YOLOv4的配电线路绝缘子与避雷器快速检测方法。针对无人机航拍的配电线路图像中各目标数目不均衡的问题,采用基于GraphCut与图像融合的样本扩充方法,建立含真实和虚拟的配电线路图像样本数据库。对于目标检测网络参数多、检测速度慢导致应用性能差的问题,通过迁移轻量级ShuffleNet网络对YOLOv4主干网络进行替换,利用深度可分离卷积原理对PANet模块进行轻量化调整 ...
王佰川, 王聪
doaj  

太阳耀斑预报深度学习建模中样本不均衡研究

open access: yesKongjian kexue xuebao
不同等级耀斑发生的频次存在数量级上的差别, 使基于常规卷积神经网络的耀斑预报模型通常难以捕捉M和X类耀斑先兆特征, 导致高等级耀斑预报精度低的问题. 本文对于这种耀斑预报中的长尾分布问题, 通过控制变量法讨论不同深度长尾学习方法对于耀斑预报精度提升. 尝试从训练集优化、损失函数优化、网络权重优化等角度改进模型对于M和X类耀斑的预报性能.
周 俊   +3 more
doaj   +1 more source

Home - About - Disclaimer - Privacy