期刊文献+
共找到4,047篇文章
< 1 2 203 >
每页显示 20 50 100
Accurate closed-form flutter eigensolutions of three-dimensional composite laminates with shear deformation
1
作者 Dezhuang PAN Yufeng XING 《Chinese Journal of Aeronautics》 2026年第1期230-246,共17页
According to the Mindlin plate theory and the first-order piston theory,this work obtains accurate closed-form eigensolutions for the flutter problem of three-dimensional(3D)rectangular laminated panels.The governing ... According to the Mindlin plate theory and the first-order piston theory,this work obtains accurate closed-form eigensolutions for the flutter problem of three-dimensional(3D)rectangular laminated panels.The governing differential equations are derived by the Hamilton's variational principle,and then solved by the iterative Separation-of-Variable(i SOV)method,which are applicable to arbitrary combinations of homogeneous Boundary Conditions(BCs).However,only the simply-support,clamped and cantilever panels are considered in this work for the sake of clarity.With the closed-form eigensolutions,the flutter frequency,flutter mode and flutter boundary are presented,and the effect of shear deformation and aerodynamic damping on flutter frequencies is investigated.Besides,the relation between panel energy and the work of aerodynamic load is discussed.The numerical comparisons reveal the following.(A)The flutter eigenvalues obtained by the present method are accurate,validated by the Finite Element Method(FEM)and the Galerkin method.(B)When the span-chord ratio is larger than 3,simplifying a 3D panel to 2D(two-dimensional)panel is reasonable and the relative differences of the flutter points predicted by the two models are less than one percent.(C)The reciprocal relationship between the mechanical energy of the panel and the work done by aerodynamic load is verified by using the present flutter eigenvalues and modes,further indicating the high accuracy of the present solutions.(D)The coupling of shear deformation and aerodynamic damping prevents frequency coalescing. 展开更多
关键词 Closed-form eigensolutions The first-order piston theory The Mindlin plate theory three-dimensional panel flutter Separation-of-variable method
原文传递
基于改进U-Net的铜合金晶界识别方法
2
作者 靖青秀 刘卫辉 +4 位作者 常琪琪 谢伟滨 张志聪 吴瑞洋 黄晓东 《有色金属(中英文)》 北大核心 2026年第2期198-206,共9页
晶粒度评级精度高度依赖于准确的晶粒尺寸与形状表征,而晶界分割是界定晶粒范围的关键预处理步骤。针对铜合金显微图像中晶界对比度低、边缘模糊导致的检测困难,以及现有高精度分割算法参数量大、计算复杂度高、难以满足工业实时检测需... 晶粒度评级精度高度依赖于准确的晶粒尺寸与形状表征,而晶界分割是界定晶粒范围的关键预处理步骤。针对铜合金显微图像中晶界对比度低、边缘模糊导致的检测困难,以及现有高精度分割算法参数量大、计算复杂度高、难以满足工业实时检测需求等问题,本文提出一种基于MobileNetV2的轻量化U-Net改进方法。通过将MobileNetV2作为主干网络解决特征丢失问题,并引入集成深度可分离卷积的ASPP模块,有效增强了多尺度语义特征提取能力。实验结果表明,改进后的模型在保持轻量化的同时,在晶界分割任务中取得了mIOU 87.66%、精确率93.50%、平均像素准确率92.79%的优异性能,显著优于传统U-Net模型,为工业现场实时晶界识别提供了可靠解决方案。 展开更多
关键词 铜合金 晶粒度 深度学习 u-net 轻量化
在线阅读 下载PDF
基于多尺度特征提取的U-Net网络微地震定位方法
3
作者 黄建平 王秋阳 +6 位作者 李媛媛 黎国龙 苏来源 路依霖 李三福 段文胜 雷刚林 《中国石油大学学报(自然科学版)》 北大核心 2026年第1期1-11,共11页
微地震定位是微地震监测的核心任务,面对当前海量的地震数据,传统的定位方法已无法满足实时定位的需求。为此,利用深度学习技术,提出一种基于U-Net网络为主要架构的微地震震源定位方法,通过融合双交叉注意力模块和空间空洞金字塔池化模... 微地震定位是微地震监测的核心任务,面对当前海量的地震数据,传统的定位方法已无法满足实时定位的需求。为此,利用深度学习技术,提出一种基于U-Net网络为主要架构的微地震震源定位方法,通过融合双交叉注意力模块和空间空洞金字塔池化模块,增强网络对微震数据中波形特征的提取能力,提升震源位置预测精度。最后,利用简单层状和复杂速度模型生成合成数据进行实验测试,并与U-Net和Att-Unet网络对震源位置预测误差精度进行对比分析。结果表明,所构建的网络模型在震源预测精度以及网络性能上均优于其他网络模型,并且对低信噪比的微地震数据也有较好的预测效果。 展开更多
关键词 微震定位 水力压裂 多尺度特征提取 u-net网络 注意力机制
在线阅读 下载PDF
基于改进U-Net网络和知识蒸馏的三维断层识别方法
4
作者 王莉利 梁云虎 高新成 《石油物探》 北大核心 2026年第1期21-30,共10页
深度学习方法在三维地震资料断层识别中得到了广泛应用,但方法的应用面临数据集质量欠佳、资源消耗过高以及训练周期长等问题。为此,提出了一种融合改进U-Net网络和知识蒸馏的三维断层识别方法。该方法先将改进的U-Net网络模型作为教师... 深度学习方法在三维地震资料断层识别中得到了广泛应用,但方法的应用面临数据集质量欠佳、资源消耗过高以及训练周期长等问题。为此,提出了一种融合改进U-Net网络和知识蒸馏的三维断层识别方法。该方法先将改进的U-Net网络模型作为教师模型,将空洞空间金字塔池化(ASPP)结构与U-Net网络模型相融合,构建轻量级学生模型,然后引入知识蒸馏技术对学生模型进行优化,并调整网络训练超参数和知识蒸馏损失参数,使学生模型获取更丰富的断层信息,提升学生模型的网络性能。该方法通过将复杂的教师模型的知识迁移到轻量级学生模型,显著降低了模型的计算复杂度,同时保持了较高的识别精度。测试结果表明,在合成测试集和实际地震数据的断层识别中,经过知识蒸馏训练的学生模型在识别精度和连续性上均优于未经过蒸馏的学生模型和单独训练的教师模型,充分验证了方法的可行性和有效性。 展开更多
关键词 断层识别 知识蒸馏 u-net 教师模型 学生模型
在线阅读 下载PDF
融合对抗自编码器和U-net的非侵入式负荷分解方法
5
作者 王凌云 朱倍萱 +1 位作者 张涛 罗明天 《电力系统及其自动化学报》 北大核心 2026年第2期59-68,共10页
为了提高非侵入式负荷分解模型的分解效果和泛化性能,并针对现有生成式模型在非侵入式负荷分解任务中存在的一些局限性,提出一种引入变分推理思想和联合对抗机制的对抗自编码器非侵入式负荷分解方法。为保证负荷分解的实时性,采用序列... 为了提高非侵入式负荷分解模型的分解效果和泛化性能,并针对现有生成式模型在非侵入式负荷分解任务中存在的一些局限性,提出一种引入变分推理思想和联合对抗机制的对抗自编码器非侵入式负荷分解方法。为保证负荷分解的实时性,采用序列到序列映射模型。基于U-net框架构建对抗自编码器模型,在编码器与解码器之间添加跳跃连接,使模型可以同时捕获电器特征的局部细节和全局信息,实现多特征融合,避免特征丢失,同时引入实例-批归一化网络,提高模型的分解性能以及泛化性能。最后将所提模型与几种代表性模型在UK-DALE数据集上进行对比实验。结果表明:所提模型具有优秀的分解性能和泛化能力,并且更加轻量化。 展开更多
关键词 非侵入式负荷分解 对抗自编码器 深度学习 序列到序列 u-net 实例-批归一化
在线阅读 下载PDF
Three-dimensional gravity inversion based on 3D U-Net++ 被引量:4
6
作者 Wang Yu-Feng Zhang Yu-Jie +1 位作者 Fu Li-Hua Li Hong-Wei 《Applied Geophysics》 SCIE CSCD 2021年第4期451-460,592,共11页
The gravity inversion is to restore genetic density distribution of the underground target to be explored for explaining the internal structure and distribution of the Earth.In this paper,we propose a new 3D gravity i... The gravity inversion is to restore genetic density distribution of the underground target to be explored for explaining the internal structure and distribution of the Earth.In this paper,we propose a new 3D gravity inversion method based on 3D U-Net++.Compared with two-dimensional gravity inversion,three-dimensional(3D)gravity inversion can more precisely describe the density distribution of underground space.However,conventional 3D gravity inversion method input is two-dimensional,the input and output of the network proposed in our method are three-dimensional.In the training stage,we design a large number of diversifi ed simulation model-data pairs by using the random walk method to improve the generalization ability of the network.In the test phase,we verify the network performance by using the model-data pairs generated by the simulation.To further illustrate the eff ectiveness of the algorithm,we apply the method to the inversion of the San Nicolas mining area,and the inversion results are basically consistent with the borehole measurement results.Moreover,the results of the 3D U-Net++inversion and the 3D U-Net inversion are compared.The density models of the 3D U-Net++inversion have higher resolution,more concentrated inversion results,and a clearer boundary of the density model. 展开更多
关键词 deep learning gravity anomaly three-dimensional gravity inversion 3D u-net++
在线阅读 下载PDF
Three-dimensional line-of-sight-angle-constrained leader-following cooperative interception guidance law with prespecified impact time 被引量:3
7
作者 Hao YOU Xinlong CHANG Jiufen ZHAO 《Chinese Journal of Aeronautics》 2025年第1期491-506,共16页
To address the problem of multi-missile cooperative interception against maneuvering targets at a prespecified impact time and desired Line-of-Sight(LOS)angles in ThreeDimensional(3D)space,this paper proposes a 3D lea... To address the problem of multi-missile cooperative interception against maneuvering targets at a prespecified impact time and desired Line-of-Sight(LOS)angles in ThreeDimensional(3D)space,this paper proposes a 3D leader-following cooperative interception guidance law.First,in the LOS direction of the leader,an impact time-controlled guidance law is derived based on the fixed-time stability theory,which enables the leader to complete the interception task at a prespecified impact time.Next,in the LOS direction of the followers,by introducing a time consensus tracking error function,a fixed-time consensus tracking guidance law is investigated to guarantee the consensus tracking convergence of the time-to-go.Then,in the direction normal to the LOS,by combining the designed global integral sliding mode surface and the second-order Sliding Mode Control(SMC)theory,an innovative 3D LOS-angle-constrained interception guidance law is developed,which eliminates the reaching phase in the traditional sliding mode guidance laws and effectively saves energy consumption.Moreover,it effectively suppresses the chattering phenomenon while avoiding the singularity issue,and compensates for unknown interference caused by target maneuvering online,making it convenient for practical engineering applications.Finally,theoretical proof analysis and multiple sets of numerical simulation results verify the effectiveness,superiority,and robustness of the investigated guidance law. 展开更多
关键词 three-dimensional cooperative interception Leader-following missiles Prespecified impact time LOS-angle-constrained Fixed-time stability Global integral sliding mode
原文传递
Global Mapping of Three-Dimensional Urban Structures Reveals Escalating Utilization in the Vertical Dimension and Pronounced Building Space Inequality 被引量:1
8
作者 Xiaoping Liu Xinxin Wu +6 位作者 Xuecao Li Xiaocong Xu Weilin Liao Limin Jiao Zhenzhong Zeng Guangzhao Chen Xia Li 《Engineering》 2025年第4期86-99,共14页
Three-dimensional(3D)urban structures play a critical role in informing climate mitigation strategies aimed at the built environment and facilitating sustainable urban development.Regrettably,there exists a significan... Three-dimensional(3D)urban structures play a critical role in informing climate mitigation strategies aimed at the built environment and facilitating sustainable urban development.Regrettably,there exists a significant gap in detailed and consistent data on 3D building space structures with global coverage due to the challenges inherent in the data collection and model calibration processes.In this study,we constructed a global urban structure(GUS-3D)dataset,including building volume,height,and footprint information,at a 500 m spatial resolution using extensive satellite observation products and numerous reference building samples.Our analysis indicated that the total volume of buildings worldwide in2015 exceeded 1×10^(12)m^(3).Over the 1985 to 2015 period,we observed a slight increase in the magnitude of 3D building volume growth(i.e.,it increased from 166.02 km3 during the 1985–2000 period to 175.08km3 during the 2000–2015 period),while the expansion magnitudes of the two-dimensional(2D)building footprint(22.51×10^(3) vs 13.29×10^(3)km^(2))and urban extent(157×10^(3) vs 133.8×10^(3)km^(2))notably decreased.This trend highlights the significant increase in intensive vertical utilization of urban land.Furthermore,we identified significant heterogeneity in building space provision and inequality across cities worldwide.This inequality is particularly pronounced in many populous Asian cities,which has been overlooked in previous studies on economic inequality.The GUS-3D dataset shows great potential to deepen our understanding of the urban environment and creates new horizons for numerous 3D urban studies. 展开更多
关键词 three-dimensional Global mapping Building volume Building height Building space inequality
在线阅读 下载PDF
Three-dimensional reconstruction under computed tomography and myopectineal orifice measurement under laparoscopy for quality control of inguinal hernia treatment 被引量:1
9
作者 Lei Zhang Jing Chen +7 位作者 Yu-Ying Zhang Lei Liu Han-Dan Wang Ya-Fei Zhang Jun Sheng Qiu-Shi Hu Ming-Liang Liu Yi-Lin Yuan 《World Journal of Gastrointestinal Endoscopy》 2025年第3期50-59,共10页
BACKGROUND Inguinal hernias are common after surgery.Tension-free repair is widely accepted as the main method for managing inguinal hernias.Adequate exposure,coverage,and repair of the myopectineal orifice(MPO)are ne... BACKGROUND Inguinal hernias are common after surgery.Tension-free repair is widely accepted as the main method for managing inguinal hernias.Adequate exposure,coverage,and repair of the myopectineal orifice(MPO)are necessary.However,due to differences in race and sex,people’s body shapes vary.According to European guidelines,the patch should measure 10 cm×15 cm.If any part of the MPO is dissected,injury to the nerves,vascular network,or organs may occur during surgery,thereby leading to inguinal discomfort,pain,and seroma formation after surgery.Therefore,accurate localization and measurement of the boundary of the MPO are crucial for selecting the optimal patch for inguinal hernia repair.AIM To compare the size of the MPO measured on three-dimensional multislice spiral computed tomography(CT)with that measured via laparoscopy and explore the relevant factors influencing the size of the MPO.METHODS Clinical data from 74 patients who underwent laparoscopic tension-free inguinal hernia repair at the General Surgery Department of the First Affiliated Hospital of Anhui University of Science and Technology between September 2022 and July 2024 were collected and analyzed retrospectively.Transabdominal preperitoneal was performed.Sixty-four males and 10 females,with an average age of 58.30±12.32 years,were included.The clinical data of the patients were collected.The boundary of the MPO was measured on three-dimensional CT images before surgery and then again during transabdominal preperitoneal.All the preoperative and intraoperative data were analyzed via paired t-tests.A t-test was used for comparisons of age,body mass index,and sex between the groups.In the comparative analysis,a P value less than 0.05 indicated a significant difference.RESULTS The boundaries of the MPO on 3-dimensional CT images measured 7.05±0.47 cm and 6.27±0.61 cm,and the area of the MPO was 19.54±3.33 cm^(2).The boundaries of the MPO during surgery were 7.18±0.51 cm and 6.17±0.40 cm.The errors were not statistically significant.However,the intraoperative BD(the width of the MPO,P=0.024,P<0.05)and preoperative AC(the length of the MPO,P=0.045,P<0.05)significantly differed according to sex.The AC and BD measurements before and during surgery were not significantly different according to age,body mass index,hernia side or hernia type(P>0.05).CONCLUSION The application of this technology can aid in determining the most appropriate dissection range and patch size. 展开更多
关键词 HERNIA INGUINAL Myopectineal orifice three-dimensional reconstruction Computed tomography Inguinal hernia
暂未订购
融合多源特征与注意力机制的改进U-Net鱼鳞坑遥感提取方法
10
作者 魏敬志 黄骁力 +4 位作者 江岭 梁明 张大鹏 王莎莎 宋音 《农业工程学报》 北大核心 2026年第2期214-224,共11页
鱼鳞坑是黄土高原典型的小型水土保持措施,由于其尺度小、分布不均,传统卫星遥感方法难以实现高精度识别。为此,该研究提出一种融合多源特征与注意力机制的深度学习鱼鳞坑遥感提取方法,构建了“特征重要性分析+注意力增强U-Net结构设计... 鱼鳞坑是黄土高原典型的小型水土保持措施,由于其尺度小、分布不均,传统卫星遥感方法难以实现高精度识别。为此,该研究提出一种融合多源特征与注意力机制的深度学习鱼鳞坑遥感提取方法,构建了“特征重要性分析+注意力增强U-Net结构设计”的技术框架。基于无人机获取的高分辨率多光谱影像与数字高程模型(digital elevation model,DEM),该研究综合运用Spearman相关系数与SHAP(Shapley additive explanations)可解释性分析方法,对光谱与地形特征进行重要性评估与冗余剔除,最终优选出4类关键特征,并据此设计了9种特征组合方案。在此基础上,采用UNet、DeepLabV3+、SegNet与FCN四种语义分割模型开展对比试验,结果表明以RGB+Slope的特征组合方案在UNet模型中识别效果最优。在模型结构方面,该研究以U-Net为基础,融合金字塔压缩注意力模块(pyramid squeeze attention module,PSAM)与多级特征注意力上采样模块(multi-scale feature attention upsampling module,MFAU),增强模型对鱼鳞坑边缘与空间结构的感知能力,并设计消融试验验证改进效果。试验结果表明,在最优特征组合的数据输入下,改进模型在测试区交并比提升2.47个百分点,F1分数提升1.34个百分点,召回率提升2.72个百分点,精确率提升1.02个百分点,表现出良好的提取精度与区域泛化能力。研究表明,特征重要性分析与注意力增强结构设计的融合策略可有效提升模型对小尺度地貌目标的识别性能,为鱼鳞坑等微地形构筑物的高精度遥感提取提供技术支撑,也为多源信息融合与深度学习模型构建提供了理论参考。 展开更多
关键词 无人机 遥感 语义分割 鱼鳞坑提取 u-net改进 注意力机制
在线阅读 下载PDF
基于改进U-Net的人工光植物工厂生菜图像分割方法
11
作者 李文豪 金文帅 +3 位作者 高晟 薛岳 毛罕平 左志宇 《农机化研究》 北大核心 2026年第6期157-163,共7页
针对植物工厂内作物在人工光环境下图像分割精度不足的问题,提出了一种改进的U-Net神经网络模型,以实现对作物图像的高精度分割。首先通过对比试验,系统比较了传统图像分割方法(如阈值、聚类和区域等)与神经网络方法在人工光环境下的分... 针对植物工厂内作物在人工光环境下图像分割精度不足的问题,提出了一种改进的U-Net神经网络模型,以实现对作物图像的高精度分割。首先通过对比试验,系统比较了传统图像分割方法(如阈值、聚类和区域等)与神经网络方法在人工光环境下的分割效果,结果表明:与传统方法相比,U-Net神经网络在分割精度和模型稳定性方面具有明显优势。然而,进一步分析U-Net模型的分割结果发现,其在复杂光照条件下的分割精度和泛化能力仍有提升空间,主要体现在边界细节处理和小目标分割的准确性不足。为此,针对性地提出3种改进策略:一是通过数据增强技术扩展训练数据集,以提升模型的鲁棒性;二是对U-Net模型的结构进行优化,改进金字塔结构以增强多尺度特征融合能力;三是采用坐标注意力机制,有效提升模型对目标区域的聚焦能力,特别是在背景复杂或光线不均的情况下。基于此,进行试验验证,结果表明:结合改进金字塔结构和坐标注意力机制的U-Net模型在分割平均精确率和平均交并比上分别达到98.19%和96.86%,相比原始U-Net模型分别提高了4.01、3.02个百分点。所提方法显著改善了人工光环境下对植物工厂作物的图像分割性能,为植物工厂内作物生长监测与精准信息采集提供了技术支持,同时为未来智能农业领域的相关研究奠定了基础。 展开更多
关键词 生菜图像分割 植物工厂 人工光 改进u-net 注意力机制 神经网络
在线阅读 下载PDF
基于改进U-Net的冷冻电镜图像去噪方法
12
作者 邓竞 曾安 金亮 《计算机应用与软件》 北大核心 2026年第1期241-248,共8页
针对冷冻电镜图像信噪比极低,并且现有去噪方法不能有效去掉复杂噪声的问题,提出一种基于改进U-Net的冷冻电镜图像去噪方法。改进方法用FCN(Fully Convolutional Networks)搭建噪声映射模块,并在原始U-Net网络中嵌入多尺度联接和宽激活... 针对冷冻电镜图像信噪比极低,并且现有去噪方法不能有效去掉复杂噪声的问题,提出一种基于改进U-Net的冷冻电镜图像去噪方法。改进方法用FCN(Fully Convolutional Networks)搭建噪声映射模块,并在原始U-Net网络中嵌入多尺度联接和宽激活密集残差块,既能提高网络的泛化能力又使模型能更好地提取和恢复特征信息,从而实现高质量的冷冻电镜图像去噪;全变差损失函数的引入用来保护输出图像中的颗粒细节信息。实验结果表明,相较于对比方法,该方法在有效去除背景噪声同时能更好地恢复颗粒细节,信噪比(Signal to Noise Ratio,SNR)也是最优,并且颗粒挑选阳性数量也得到提升。 展开更多
关键词 图像去噪 冷冻电镜 u-net FCN
在线阅读 下载PDF
基于U-Net的花生网纹分割与品种识别
13
作者 巩秀钇 踪姿艳 +6 位作者 付华宇 张贺 纪翔 朱春雨 王聪 赵延伸 韩仲志 《花生学报》 北大核心 2026年第1期23-33,共11页
花生是我国重要的油料作物,不同品种花生在生长特性、产量潜力和抗逆性等方面存在显著差异。网纹作为花生荚果的独特纹理特征,在形态、密度和分布上具有显著的品种特异性,是DUS测试的重要荚果性状,但现有研究对此利用不足。因此,本研究... 花生是我国重要的油料作物,不同品种花生在生长特性、产量潜力和抗逆性等方面存在显著差异。网纹作为花生荚果的独特纹理特征,在形态、密度和分布上具有显著的品种特异性,是DUS测试的重要荚果性状,但现有研究对此利用不足。因此,本研究提出基于U-Net模型的花生网纹分割与多模态特征融合的品种识别框架。U-Net模型在对13个花生品种的网纹分割任务中表现优异,平均交并比为75.9%、准确率为89.2%,显著优于其他现有基础模型。进一步提取网纹图像的16个PCA降维特征,结合形态与颜色特征构建多模态数据集,采用SVM分类器实现品种识别,准确率达90.15%,较花生纹理、形态和颜色特征结合提升4.44%。研究首次证实花生网纹作为DUS测试性状的有效性,突破传统形态学的分析局限,为花生表型组学研究提供了可解释的方法,对推动精准育种和种质资源保护具有重要意义。 展开更多
关键词 花生网纹 DUS性状 u-net 图像分割 品种识别
在线阅读 下载PDF
Artificial intelligence-aided semi-automatic joint trace detection from textured three-dimensional models of rock mass 被引量:1
14
作者 Seyedahmad Mehrishal Jineon Kim +1 位作者 Yulong Shao Jae Joon Song 《Journal of Rock Mechanics and Geotechnical Engineering》 2025年第4期1973-1985,共13页
It is of great importance to obtain precise trace data,as traces are frequently the sole visible and measurable parameter in most outcrops.The manual recognition and detection of traces on high-resolution three-dimens... It is of great importance to obtain precise trace data,as traces are frequently the sole visible and measurable parameter in most outcrops.The manual recognition and detection of traces on high-resolution three-dimensional(3D)models are relatively straightforward but time-consuming.One potential solution to enhance this process is to use machine learning algorithms to detect the 3D traces.In this study,a unique pixel-wise texture mapper algorithm generates a dense point cloud representation of an outcrop with the precise resolution of the original textured 3D model.A virtual digital image rendering was then employed to capture virtual images of selected regions.This technique helps to overcome limitations caused by the surface morphology of the rock mass,such as restricted access,lighting conditions,and shading effects.After AI-powered trace detection on two-dimensional(2D)images,a 3D data structuring technique was applied to the selected trace pixels.In the 3D data structuring,the trace data were structured through 2D thinning,3D reprojection,clustering,segmentation,and segment linking.Finally,the linked segments were exported as 3D polylines,with each polyline in the output corresponding to a trace.The efficacy of the proposed method was assessed using a 3D model of a real-world case study,which was used to compare the results of artificial intelligence(AI)-aided and human intelligence trace detection.Rosette diagrams,which visualize the distribution of trace orientations,confirmed the high similarity between the automatically and manually generated trace maps.In conclusion,the proposed semi-automatic method was easy to use,fast,and accurate in detecting the dominant jointing system of the rock mass. 展开更多
关键词 Automatic trace detection Digital joint mapping Rock discontinuities characterization three-dimensional(3D)trace network
在线阅读 下载PDF
Dynamic modeling of a three-dimensional braided composite thin plate considering braiding directions 被引量:1
15
作者 Chentong GAO Huiyu SUN +1 位作者 Jianping GU W.M.HUANG 《Applied Mathematics and Mechanics(English Edition)》 2025年第1期123-138,共16页
Currently,there are a limited number of dynamic models available for braided composite plates with large overall motions,despite the incorporation of three-dimensional(3D)braided composites into rotating blade compone... Currently,there are a limited number of dynamic models available for braided composite plates with large overall motions,despite the incorporation of three-dimensional(3D)braided composites into rotating blade components.In this paper,a dynamic model of 3D 4-directional braided composite thin plates considering braiding directions is established.Based on Kirchhoff's plate assumptions,the displacement variables of the plate are expressed.By incorporating the braiding directions into the constitutive equation of the braided composites,the dynamic model of the plate considering braiding directions is obtained.The effects of the speeds,braiding directions,and braided angles on the responses of the plate with fixed-axis rotation and translational motion,respectively,are investigated.This paper presents a dynamic theory for calculating the deformation of 3D braided composite structures undergoing both translational and rotational motions.It also provides a simulation method for investigating the dynamic behavior of non-isotropic material plates in various applications. 展开更多
关键词 three-dimensional(3D)braided composite braiding direction composite thin plate large overall motion dynamic model
在线阅读 下载PDF
DBU-Net:基于改进SAM的双分支U-Net图像隐写方法
16
作者 周倩 朱梦月 王帆 《南京邮电大学学报(自然科学版)》 北大核心 2026年第1期124-134,共11页
基于深度学习的图像隐写术通过引入空间注意力机制使得模型关注图像的重要区域,但主流方法通常依赖固定结构的卷积模块生成单一的注意力权重矩阵,对局部细节特征关注度不足,导致信息隐藏未能与纹理复杂度充分关联。针对上述问题,提出了... 基于深度学习的图像隐写术通过引入空间注意力机制使得模型关注图像的重要区域,但主流方法通常依赖固定结构的卷积模块生成单一的注意力权重矩阵,对局部细节特征关注度不足,导致信息隐藏未能与纹理复杂度充分关联。针对上述问题,提出了一种基于改进空间注意力机制的双分支U-Net图像隐写方法DBU-Net。该方法首先引入改进的空间注意力机制对载体图像进行区域划分。其次采用差异化的信息隐藏,其中,高纹理分支采用更深层次和多尺度的卷积模块进行特征提取,同时在特征融合时被赋予更高权重,低纹理分支则采用浅层和固定卷积进行特征提取,且在特征融合时权重较低。在不同数据集上的大量对照实验结果表明,相较于现有的先进图像隐写方法,所提的DBU-Net在隐写质量、隐写不可见性和安全性方面均取得了显著的性能提升。 展开更多
关键词 图像隐写术 u-net 分支网络 空间注意力机制
在线阅读 下载PDF
Three-dimensional models:from cell culture to Patient-Derived Organoid and its application to future liposarcoma research
17
作者 SAYUMI TAHARA SYDNEY RENTSCH +4 位作者 FERNANDA COSTAS CASAL DE FARIA PATRICIA SARCHET ROMA KARNA FEDERICA CALORE RAPHAEL E.POLLOCK 《Oncology Research》 SCIE 2025年第1期1-13,共13页
Liposarcoma is one of the most common soft tissue sarcomas,however,its occurrence rate is still rare compared to other cancers.Due to its rarity,in vitro experiments are an essential approach to elucidate liposarcoma ... Liposarcoma is one of the most common soft tissue sarcomas,however,its occurrence rate is still rare compared to other cancers.Due to its rarity,in vitro experiments are an essential approach to elucidate liposarcoma pathobiology.Conventional cell culture-based research(2D cell culture)is still playing a pivotal role,while several shortcomings have been recently under discussion.In vivo,mouse models are usually adopted for pre-clinical analyses with expectations to overcome the issues of 2D cell culture.However,they do not fully recapitulate human dedifferentiated liposarcoma(DDLPS)characteristics.Therefore,three-dimensional(3D)culture systems have been the recent research focus in the cell biology field with the expectation to overcome at the same time the disadvantages of 2D cell culture and in vivo animal models and fill in the gap between them.Given the liposarcoma rarity,we believe that 3D cell culture techniques,including 3D cell cultures/co-cultures,and Patient-Derived tumor Organoids(PDOs),represent a promising approach to facilitate liposarcoma investigation and elucidate its molecular mechanisms and effective therapy development.In this review,we first provide a general overview of 3D cell cultures compared to 2D cell cultures.We then focus on one of the recent 3D cell culture applications,Patient-Derived Organoids(PDOs),summarizing and discussing several PDO methodologies.Finally,we discuss the current and future applications of PDOs to sarcoma,particularly in the field of liposarcoma. 展开更多
关键词 Cell culture LIPOSARCOMA Patient-Derived Organoid(PDO) SPHEROID three-dimensional(3D)cell culture
暂未订购
基于U-Net架构和无人机航拍传感器的公路图像裂缝检测
18
作者 陈巍 陈恳 朱文耀 《传感器与微系统》 北大核心 2026年第2期161-166,共6页
针对当前模型对公路裂缝检测不精确的问题,提出了一种基于无人机(UAV)航拍传感器遥感图像的智能检测方法。基于U-Net架构,结合深度可分离残差块(DR-Block)、空间金字塔融合注意力模块(SPFAM)和感受野块(RFB),提出DAR-Unet逐像素裂缝检... 针对当前模型对公路裂缝检测不精确的问题,提出了一种基于无人机(UAV)航拍传感器遥感图像的智能检测方法。基于U-Net架构,结合深度可分离残差块(DR-Block)、空间金字塔融合注意力模块(SPFAM)和感受野块(RFB),提出DAR-Unet逐像素裂缝检测模型。利用无人机采集1 046张高质量公路遥感图像构建专用数据集。在自制数据集上,DAR-Unet的平均交并比(mIoU)和F1分数分别达到76.41%和74.24%,高于主流模型。进一步将模型与无人机集成,构建了公路裂缝检测物联网系统,实际测试表现优异,验证了DAR-Unet在遥感图像公路裂缝检测中的有效性。 展开更多
关键词 无人机 航拍传感器 遥感图像 公路裂缝检测 u-net架构
在线阅读 下载PDF
基于改进U-Net与RGB-D图像的青花椒枝条“下桩”剪切点定位
19
作者 蒲应俊 张文州 +3 位作者 李金广 赵立军 陈子文 杨明金 《农业工程学报》 北大核心 2026年第1期160-170,共11页
青花椒枝条“下桩”是通过剪下带鲜果的枝条并保留一定长度短桩的采摘收获方法。为实现青花椒采摘机器人精准识别枝条并确定最佳剪切点以达到高效“下桩”作业,该研究提出了一种基于U-Net深度学习网络和RGB-D相机相结合的青花椒主枝“... 青花椒枝条“下桩”是通过剪下带鲜果的枝条并保留一定长度短桩的采摘收获方法。为实现青花椒采摘机器人精准识别枝条并确定最佳剪切点以达到高效“下桩”作业,该研究提出了一种基于U-Net深度学习网络和RGB-D相机相结合的青花椒主枝“下桩”剪切点定位方法。首先,通过改进传统U-Net模型,将其主干网络替换为嵌入CA注意力机制的ResNet50网络,同时在U-Net模型的特征拼接阶段中增加SE注意力机制,从而构建针对青花椒主枝和树干的分割模型。然后,将分割后的图像利用二值化与骨架线提取方法得到主枝中心线,结合RGB-D相机的深度信息与OpenCV图像处理算法,完成世界坐标系与像素坐标系间长度的映射。随后,将短桩预设的40 mm长度从世界坐标系映射至RGB图像中的像素长度,最终确定每根主枝的“下桩”剪切点位置。试验结果表明,改进后的U-Net模型在分割性能上优于DeeplabV3+和PSPNet,平均交并比(MIoU)、平均像素准确率(mPA)和召回率(recall)分别达到87.58%、93.76%和96.24%。在晴天顺光、逆光及阴天条件下,“下桩”剪切点识别定位的成功率分别达到90.81%、84.88%、80.52%。采摘点定位试验中,定位成功率为90%,单根花椒枝平均识别过程耗时1.93 s。该研究结果可为青花椒采摘机器人“下桩”采收提供技术支撑。 展开更多
关键词 图像处理 青花椒 采摘 u-net网络模型 下桩采摘法 剪切点定位
在线阅读 下载PDF
基于多通道U-Net的室内电磁波传播路径损耗预测方法
20
作者 关策 宋欣蔚 岳云涛 《电波科学学报》 北大核心 2026年第1期79-88,共10页
精准高效的室内电磁波传播预测是无线通信系统规划与优化的关键基础。传统的方法在高频段因电磁波传播特性复杂、计算复杂度剧增,难以满足实时性需求;而现有深度学习模型在高频段适配性与预测精度上也存在一定的局限性。本文提出一种多... 精准高效的室内电磁波传播预测是无线通信系统规划与优化的关键基础。传统的方法在高频段因电磁波传播特性复杂、计算复杂度剧增,难以满足实时性需求;而现有深度学习模型在高频段适配性与预测精度上也存在一定的局限性。本文提出一种多通道U-Net的深度学习预测模型(mU-Net),其将天线方向特性、反射通道贡献、透射通道贡献及自由空间路径损耗(free space path loss, FSPL)分别建模为4个独立特征通道,通过多通道输入U-Net网络捕捉不同传播机制的差异化影响,mU-Net非对称编码器-解码器结构进一步融合多源特征,解析复杂结构散射与绕射细节,从而输出高分辨率路径损耗(path loss, PL)预测图。该模型能深度融合高频段特征信息,解决了现有方法实时性差、精度不足的问题。3种场景下的PL预测结果显示,所提出的mU-Net模型在精度、结构相似性及预测效率等关键指标上均显著优于现有的典型预测方法,在高频段场景下性能提升尤为明显,为高频段室内电磁波传播预测提供了兼具高精度与高效率的解决方案。 展开更多
关键词 室内电磁波传播 路径损耗(PL) 高频段 深度学习 多通道 u-net
在线阅读 下载PDF
上一页 1 2 203 下一页 到第
使用帮助 返回顶部