期刊文献+
共找到497篇文章
< 1 2 25 >
每页显示 20 50 100
Multi-layer Tectonic Model for Intraplate Deformation and Plastic-Flow Network in the Asian Continental Lithosphere 被引量:4
1
作者 Wang Shengzu Institute of Geology, State Seismological Bureau, Beijing Liu Linqun 《Acta Geologica Sinica(English Edition)》 SCIE CAS CSCD 1993年第3期247-271,共25页
In a large area of the east—central Asian continent there is a unified seismic network system composed of two families of large—seismic belts that intersect conjugately. Such a seismic network in the middle—upper c... In a large area of the east—central Asian continent there is a unified seismic network system composed of two families of large—seismic belts that intersect conjugately. Such a seismic network in the middle—upper crust is actually a response to the plastic flow network in the lower lithosphere including the lower crust and lithospheric mantle. The existence of the unified plastic flow system confirms that the driving force for intraplate tectonic deformation results mainly from the compression of the India plate, while the long-range transmission of the force is carried out chiefly by means of plastic flow. The plastic flow network has a control over the intraplate tectonic deformation. 展开更多
关键词 Continental lithosphere tectonic deformation multi-layer tectonic model large-scale seismic belt seismic network plastic flow network
下载PDF
Explosive synchronization of multi-layer complex networks based on star connection between layers with delay
2
作者 金彦亮 韩钱源 +2 位作者 郭润珠 高塬 沈礼权 《Chinese Physics B》 SCIE EI CAS CSCD 2023年第10期343-349,共7页
Explosive synchronization(ES)is a kind of first-order jump phenomenon that exists in physical and biological systems.In recent years,researchers have focused on ES between single-layer and multi-layer networks.Most re... Explosive synchronization(ES)is a kind of first-order jump phenomenon that exists in physical and biological systems.In recent years,researchers have focused on ES between single-layer and multi-layer networks.Most research on complex networks with delay has focused on single-layer or double-layer networks,multi-layer networks are seldom explored.In this paper,we propose a Kuramoto model of frequency weights in multi-layer complex networks with delay and star connections between layers.Through theoretical analysis and numerical verification,the factors affecting the backward critical coupling strength are analyzed.The results show that the interaction between layers and the average node degree has a direct effect on the backward critical coupling strength of each layer network.The location of the delay,the size of the delay,the number of network layers,the number of nodes,and the network topology are revealed to have no direct impact on the backward critical coupling strength of the network.Delay is introduced to explore the influence of delay and other related parameters on ES. 展开更多
关键词 multi-layer networks Kuramoto model explosive synchronization DELAY
下载PDF
Prediction of coal ash fusion temperatures using computational intelligence based models 被引量:3
3
作者 Sanjeev S.Tambe Makarand Naniwadekar +2 位作者 Shishir Tivvary Ashis Mukherjee Tarit Baran Das 《International Journal of Coal Science & Technology》 EI 2018年第4期486-507,共22页
In the coal-based combustion and gasification processes, the mineral matter contained in the coal (predominantly oxides), is left as an incombustible residue, termed ash. Commonly, ash deposits are formed on the heat ... In the coal-based combustion and gasification processes, the mineral matter contained in the coal (predominantly oxides), is left as an incombustible residue, termed ash. Commonly, ash deposits are formed on the heat absorbing surfaces of the exposed equipment of the combustion/gasification processes. These deposits lead to the occurrence of slagging or fouling and. consequently, reduced process efficiency. The ash fusion temperatures (AFTs) signify the temperature range over which the ash deposits are formed on the heat absorbing surfaces of the process equipment. Thus, for designing and operating the coal-based processes, it is important to have mathematical models predicting accurately the four types of AFTs namely initial deformation temperature, softening temperature, hemispherical temperature, and flow temperature. Several linear/nonlinear models with varying prediction accuracies and complexities are available for the AFT prediction. Their principal drawback is their applicability to the coals originating from a limited number of geographical regions. Accordingly, this study presents computational intelligenee (CI) based nonlinear models to predict the four AFTs using the oxide composition of the coal ash as the model input. The CI methods used in the modeling are genetic programming (GP), artificial neural networks, and support vector regression. The no table features of this study are that the models with a better AFT prediction and generalization performanee, a wider application potential, and reduced complexity, have been developed. Among the Ci-based models, GP and MLP based models have yielded overall improved performanee in predicting all four AFTs. 展开更多
关键词 ASH fusion temperature Artificial neural networks Support VECTOR regression GENETIC PROGRAMMING DATA-DRIVEN modeling
下载PDF
Explosive synchronization of multi-layer complex networks based on inter-layer star network connection
4
作者 Yan-Liang Jin Run-Zhu Guo +1 位作者 Xiao-Qi Yu Li-Quan Shen 《Chinese Physics B》 SCIE EI CAS CSCD 2021年第12期264-270,共7页
Explosive synchronization(ES)is a first-order transition phenomenon that is ubiquitous in various physical and biological systems.In recent years,researchers have focused on explosive synchronization in a single-layer... Explosive synchronization(ES)is a first-order transition phenomenon that is ubiquitous in various physical and biological systems.In recent years,researchers have focused on explosive synchronization in a single-layer network,but few in multi-layer networks.This paper proposes a frequency-weighted Kuramoto model in multi-layer complex networks with star connection between layers and analyzes the factors affecting the backward critical coupling strength by both theoretical analysis and numerical validation.Our results show that the backward critical coupling strength of each layer network is influenced by the inter-layer interaction strength and the average degree.The number of network layers,the number of nodes,and the network topology can not directly affect the synchronization of the network.Enhancing the inter-layer interaction strength can prevent the emergence of explosive synchronization and increasing the average degree can promote the generation of explosive synchronization. 展开更多
关键词 explosive synchronization Kuramoto model multi-layer networks
下载PDF
Action Recognition in Surveillance Videos with Combined Deep Network Models
5
作者 ZHANG Diankai ZHAO Rui-Wei +3 位作者 SHEN Lin CHEN Shaoxiang SUN Zhenfeng JIANG Yu-Gang 《ZTE Communications》 2016年第B12期54-60,共7页
Action recognition is an important topic in computer vision. Recently, deep learning technologies have been successfully used in lots of applications including video data for sloving recognition problems. However, mos... Action recognition is an important topic in computer vision. Recently, deep learning technologies have been successfully used in lots of applications including video data for sloving recognition problems. However, most existing deep learning based recognition frameworks are not optimized for action in the surveillance videos. In this paper, we propose a novel method to deal with the recognition of different types of actions in outdoor surveillance videos. The proposed method first introduces motion compensation to improve the detection of human target. Then, it uses three different types of deep models with single and sequenced images as inputs for the recognition of different types of actions. Finally, predictions from different models are fused with a linear model. Experimental results show that the proposed method works well on the real surveillance videos. 展开更多
关键词 action recognition deep network models model fusion surveillance video
下载PDF
Multimodality image registration and fusion using neural network
6
作者 Mostafa G Mostafa Aly A Farag Edward Essock 《Journal of Harbin Institute of Technology(New Series)》 EI CAS 2003年第3期235-240,共6页
Multimodality image registration and fusion are essential steps in building 3-D models from remotesensing data. We present in this paper a neural network technique for the registration and fusion of multimodali-ty rem... Multimodality image registration and fusion are essential steps in building 3-D models from remotesensing data. We present in this paper a neural network technique for the registration and fusion of multimodali-ty remote sensing data for the reconstruction of 3-D models of terrain regions. A FeedForward neural network isused to fuse the intensity data sets with the spatial data set after learning its geometry. Results on real data arepresented. Human performance evaluation is assessed on several perceptual tests in order to evaluate the fusionresults. 展开更多
关键词 data fusion image registration image interpolation neural network 3-D model building
下载PDF
Comparative Appraisal of Response Surface Methodology and Artificial Neural Network Method for Stabilized Turbulent Confined Jet Diffusion Flames Using Bluff-Body Burners
7
作者 Tahani S. Gendy Salwa A. Ghoneim Amal S. Zakhary 《World Journal of Engineering and Technology》 2020年第1期121-143,共23页
The present study was conducted to present the comparative modeling, predictive and generalization abilities of response surface methodology (RSM) and artificial neural network (ANN) for the thermal structure of stabi... The present study was conducted to present the comparative modeling, predictive and generalization abilities of response surface methodology (RSM) and artificial neural network (ANN) for the thermal structure of stabilized confined jet diffusion flames in the presence of different geometries of bluff-body burners. Two stabilizer disc burners tapered at 30° and 60° and another frustum cone of 60°/30° inclination angle were employed all having the same diameter of 80 (mm) acting as flame holders. The measured radial mean temperature profiles of the developed stabilized flames at different normalized axial distances (x/dj) were considered as the model example of the physical process. The RSM and ANN methods analyze the effect of the two operating parameters namely (r), the radial distance from the center line of the flame, and (x/dj) on the measured temperature of the flames, to find the predicted maximum temperature and the corresponding process variables. A three-layered Feed Forward Neural Network in conjugation with the hyperbolic tangent sigmoid (tansig) as transfer function and the optimized topology of 2:10:1 (input neurons: hidden neurons: output neurons) was developed. Also the ANN method has been employed to illustrate such effects in the three and two dimensions and shows the location of the predicted maximum temperature. The results indicated the superiority of ANN in the prediction capability as the ranges of R2 and F Ratio are 0.868 - 0.947 and 231.7 - 864.1 for RSM method compared to 0.964 - 0.987 and 2878.8 7580.7 for ANN method beside lower values for error analysis terms. 展开更多
关键词 STABILIZED TURBULENT Flames BLUFF-BODY Burners Thermal Structure modeling Artificial NEURAL network Response Surface Methodology multi-layer PERCEPTRON Feed Forward NEURAL network
下载PDF
Multi-dimensional Simulation of Phase Change by a 0D-2D Model Coupling via Stefan Condition
8
作者 Adrien Drouillet Romain Le Tellier +2 位作者 Raphaël Loubère Mathieu Peybernes Louis Viot 《Communications on Applied Mathematics and Computation》 2023年第2期853-884,共32页
Considering phase changes associated with a high-temperature molten material cooled down from the outside,this work presents an improvement of the modelling and the numerical simulation of such processes for an applic... Considering phase changes associated with a high-temperature molten material cooled down from the outside,this work presents an improvement of the modelling and the numerical simulation of such processes for an application pertaining to the safety of light water nuclear reactors.Postulating a core meltdown accident,the behaviour of the core melt(aka corium)into a steel vessel is of tremendous importance when evaluating the vessel integrity.Evaluating correctly the heat fluxes requires the numerical simulation of the interaction between the liquid material and its solid counterpart which forms during the solidification process,but also may melt back.To simulate this configuration,encoun-tered in various industrial applications,one considers a bi-phase model constituted by a liquid phase in contact and interaction with its solid phase.The liquid phase may solidify in presence of low energetic source,while the solid phase may melt due to an intense heat flux from the high-temperature liquid.In the frame of the in-house legacy code,several simplifying assumptions(0D multi-layer discretization,instantaneous heat transfer via a quadratic temperature profile in solids)are made for the modelling of such phase changes.In the present work,these shortcomings are illustrated and further overcome by solving a 2D heat conduction model in the solid by a mixed Raviart-Thomas finite element method coupled to the liquid phase due to heat and mass exchanges through Stefan condition.The liquid phase is modeled with a 0D multi-layer approach.The 0D-liquid and 2D-solid mod-els are coupled by a Stefan like phase change interface model.Several sanity checks are performed to assess the validity of the approach on 1D and 2D academical configurations for which exact or reference solutions are available.Then more advanced situations(genu-ine multi-dimensional phase changes and an"industrial-like scenario")are simulated to verify the appropriate behavior of the obtained coupled simulation scheme. 展开更多
关键词 Simulation of phase change fusion SOLIDIFICATION 0D multi-layer model 2D heat conduction model model coupling
下载PDF
Application Research on Two-Layer Threat Prediction Model Based on Event Graph
9
作者 Shuqin Zhang Xinyu Su +2 位作者 Yunfei Han Tianhui Du Peiyu Shi 《Computers, Materials & Continua》 SCIE EI 2023年第12期3993-4023,共31页
Advanced Persistent Threat(APT)is now the most common network assault.However,the existing threat analysis models cannot simultaneously predict the macro-development trend and micro-propagation path of APT attacks.The... Advanced Persistent Threat(APT)is now the most common network assault.However,the existing threat analysis models cannot simultaneously predict the macro-development trend and micro-propagation path of APT attacks.They cannot provide rapid and accurate early warning and decision responses to the present system state because they are inadequate at deducing the risk evolution rules of network threats.To address the above problems,firstly,this paper constructs the multi-source threat element analysis ontology(MTEAO)by integrating multi-source network security knowledge bases.Subsequently,based on MTEAO,we propose a two-layer threat prediction model(TL-TPM)that combines the knowledge graph and the event graph.The macro-layer of TL-TPM is based on the knowledge graph to derive the propagation path of threats among devices and to correlate threat elements for threat warning and decision-making;The micro-layer ingeniously maps the attack graph onto the event graph and derives the evolution path of attack techniques based on the event graph to improve the explainability of the evolution of threat events.The experiment’s results demonstrate that TL-TPM can completely depict the threat development trend,and the early warning results are more precise and scientific,offering knowledge and guidance for active defense. 展开更多
关键词 Knowledge graph multi-source data fusion network security threat modeling event graph absorbing Markov chain threat propagation path
下载PDF
支撑新型配电网数字化规划的图形⁃模型⁃数据融合关键技术 被引量:3
10
作者 余涛 王梓耀 +3 位作者 孙立明 曹华珍 吴亚雄 吴毓峰 《电力系统自动化》 EI CSCD 北大核心 2024年第6期139-153,共15页
配电网规划领域期盼实现智能规划,其愿景在于实现无人或少人干预的全自动规划。在数字化转型的背景下,新型配电网规划将面临图形多样化、场景碎片化、数据规模化三大挑战。文中从图形-模型-数据融合的角度提出三大关键技术:基于电气图... 配电网规划领域期盼实现智能规划,其愿景在于实现无人或少人干预的全自动规划。在数字化转型的背景下,新型配电网规划将面临图形多样化、场景碎片化、数据规模化三大挑战。文中从图形-模型-数据融合的角度提出三大关键技术:基于电气图纸识别和拓扑智能分析的图形-模型融合技术、基于知识驱动的负荷/新能源推演分析和智能决策的模型-数据融合技术、基于多模态数据融合和多时空数据联动的图形-数据融合技术,尝试打破理论研究与数字化工程的壁垒。最后,对未来新型配电网数字化规划的发展进行思考和展望,为实现“以机为主,人机协同”的大闭环模式提供借鉴。 展开更多
关键词 图形-模型-数据融合 配电网 数字化规划 知识驱动 图计算
下载PDF
结合轻量化与多尺度融合的交通标志检测算法 被引量:1
11
作者 兰红 王惠钊 《计算机工程》 CAS CSCD 北大核心 2024年第10期381-392,共12页
交通标志检测在自动驾驶领域具有重要的应用价值,及时准确地检测交通目标对提高驾驶安全性和预防交通事故具有重要意义。针对交通标志尺寸小,易受遮挡,在复杂环境下容易出现漏检、错检等问题,在YOLOv8的结构基础上提出一种结合轻量化与... 交通标志检测在自动驾驶领域具有重要的应用价值,及时准确地检测交通目标对提高驾驶安全性和预防交通事故具有重要意义。针对交通标志尺寸小,易受遮挡,在复杂环境下容易出现漏检、错检等问题,在YOLOv8的结构基础上提出一种结合轻量化与多尺度融合的交通标志检测网络架构M-YOLO,构建M-YOLOs模型来应对高精度需求的检测任务,并调整网络深度得到更轻量化的M-YOLOn模型来解决不同环境下的检测需求。首先针对交通标志目标尺寸小、图像特征流失的问题,通过增加小目标检测层,保留更多的特征信息,提高网络对于小目标的特征学习能力。提出高效多尺度特征金字塔融合网络MPANet,将浅层特征图进行降维与跳跃连接,从而融合更多的图像特征信息。然后提出融合稀疏注意力和空间注意力的BRSA注意力模块,有效提取全局和局部的位置信息,减少复杂背景下对于关键信息的干扰。最后设计两种轻量高效的BBot模块和C2fGhost模块,以提高模型运算速度并减少参数量。实验结果表明,M-YOLO相较于YOLOv8,参数量降低约1/3。在TT100K数据集和GTSDB数据集上,M-YOLOs检测精度分别提升了9.7和2.1个百分点,M-YOLOn检测精度分别提升了14.5和2.6个百分点,在轻量化的同时具备更高的检测效果。M-YOLO架构解决了浅层特征图在特征提取过程中信息丢失的问题,并显著降低模型特征提取过程中冗余的计算开销,在实景采集的数据集上证实效果有效,表明在交通标志检测任务中具有应用价值。 展开更多
关键词 卷积神经网络 轻量化模型 目标检测 注意力模块 多尺度融合
下载PDF
考虑数据不足和基于合作博弈模型融合的风电机组轴承故障诊断方法
12
作者 李俊卿 胡晓东 +2 位作者 王罗 马亚鹏 何玉灵 《太阳能学报》 EI CAS CSCD 北大核心 2024年第1期234-241,共8页
针对风电机组轴承疲劳实验成本高导致故障数据不足的问题,提出基于粒子群算法(PSO)优化的辅助分类器生成对抗网络(ACGAN),利用PSO对ACGAN的参数进行寻优,进而利用ACGAN生成与原始样本高度相似的新样本;针对单一模型对风电机组轴承故障... 针对风电机组轴承疲劳实验成本高导致故障数据不足的问题,提出基于粒子群算法(PSO)优化的辅助分类器生成对抗网络(ACGAN),利用PSO对ACGAN的参数进行寻优,进而利用ACGAN生成与原始样本高度相似的新样本;针对单一模型对风电机组轴承故障诊断的准确率较低的缺点,引进合作博弈理论对多个子模型的诊断结果进行融合,将各个子模型的诊断概率矩阵由合作博弈理论进行融合并输出最终的诊断结果。实验证明,优化后的ACGAN模型和基于合作博弈的模型融合能有效提高轴承故障诊断的准确率。 展开更多
关键词 风电机组 轴承 生成式对抗网络 故障诊断 模型融合 合作博弈
下载PDF
基于残差密集融合对抗生成网络的PET-MRI图像融合
13
作者 刘尚旺 杨荔涵 《河南师范大学学报(自然科学版)》 CAS 北大核心 2024年第1期74-83,I0005,共11页
为了增强核磁共振与正电子发射断层扫描图像融合的纹理细节,摆脱人工设计融合规则对先验知识的依赖.提出了自适应的残差密集生成对抗网络(adaptive dense residual generative adversarial network,ADRGAN)来融合两种模态的医学图像.ADR... 为了增强核磁共振与正电子发射断层扫描图像融合的纹理细节,摆脱人工设计融合规则对先验知识的依赖.提出了自适应的残差密集生成对抗网络(adaptive dense residual generative adversarial network,ADRGAN)来融合两种模态的医学图像.ADRGAN设计了区域残差学习模块与输出级联生成器,在加深网络结构的同时避免特征丢失;然后,设计了基于自适应模块的内容损失函数,强化输出融合图像的内容信息;最后,通过源图像的联合梯度图与融合图像的梯度图构建对抗性博弈来高效训练生成器与鉴别器.实验结果表明,ADRGAN在哈佛医学院MRI/PET数据集的测试中峰值信噪比和结构相似度分别达到55.2124和0.4697,均优于目前最先进的算法;所构建的模型具有端对端和无监督两特性,无需人工干预,也不需要真实数据作为标签. 展开更多
关键词 深度学习 对抗生成网络 多模态图像融合 密集残差网络
下载PDF
基于文本和声学特征的双模态融合抑郁倾向识别算法
14
作者 赵健 崔骞 +1 位作者 石佳 刘岳 《计算机工程》 CAS CSCD 北大核心 2024年第11期49-58,共10页
在抑郁症诊断中,抑郁症患者的面部表情、声音信号和文字等数据可以作为评估抑郁倾向的客观指标。相较于视频,文本和音频模态在处理敏感的个人信息时能更好地保护患者的隐私,并且文本和音频均属于语言模态,相关性较强。针对抑郁倾向识别... 在抑郁症诊断中,抑郁症患者的面部表情、声音信号和文字等数据可以作为评估抑郁倾向的客观指标。相较于视频,文本和音频模态在处理敏感的个人信息时能更好地保护患者的隐私,并且文本和音频均属于语言模态,相关性较强。针对抑郁倾向识别中变长文本数据不易被分析以及手动提取音频特征存在局限性的问题,提出一种基于Transformer的融合网络优化方法。对于文本模态,使用卷积神经网络对文本进行特征提取,得到文本在不同尺度下的局部特征,然后引入Transformer模型来处理全局信息和长距离依赖。对于音频模态,为了降低手动提取音频特征对识别结果的影响,通过使用VGGish网络来自动提取音频特征,并将提取好的音频特征送入Transformer中。最后,为进一步增强文本和音频模态融合网络的识别性能,引入SE通道注意力机制,使模型能够自适应地调整各模态之间的权重分配,更有效地聚焦于关键特征。实验结果表明,双模态融合后的网络准确率达到92.7%,相比仅使用文本或音频模态,准确率分别提升2.9和4.9个百分点。 展开更多
关键词 Transformer模型 VGGish网络 双模态融合 抑郁倾向识别 SE通道注意力机制 深度学习
下载PDF
基于DS证据理论的电网信息自动融合模型构建
15
作者 汤德荣 《佳木斯大学学报(自然科学版)》 CAS 2024年第7期52-55,共4页
多智能体感知的智能电网信息中,通常存在信息冗余与缺失等问题。为解决这些问题,研究构建智能电网多智能体信息自动化融合模型,精准自动化融合多智能体信息,改进DS证据组合规则,完成决策级信息自动化融合,提升自动化融合效果,为智能电... 多智能体感知的智能电网信息中,通常存在信息冗余与缺失等问题。为解决这些问题,研究构建智能电网多智能体信息自动化融合模型,精准自动化融合多智能体信息,改进DS证据组合规则,完成决策级信息自动化融合,提升自动化融合效果,为智能电网的后续应用,提供更加全面的信息。 展开更多
关键词 智能电网 多智能体信息 自动化 融合模型 神经网络
下载PDF
基于实测数据融合的堆芯物理模型反演优化方法及工业验证研究 被引量:1
16
作者 郭林 张凯 +1 位作者 万承辉 吴宏春 《原子能科学技术》 EI CAS CSCD 北大核心 2024年第7期1432-1439,共8页
由于堆芯运行过程中的组件辐照生长、冷却剂高速冲击等因素,燃料组件不可避免地会出现弯曲现象。但机组运行期间无法直接测量燃料组件弯曲状态,导致数值模拟采用的堆芯物理模型与真实堆芯状态之间存在差异,直观上表现为堆芯功率分布的... 由于堆芯运行过程中的组件辐照生长、冷却剂高速冲击等因素,燃料组件不可避免地会出现弯曲现象。但机组运行期间无法直接测量燃料组件弯曲状态,导致数值模拟采用的堆芯物理模型与真实堆芯状态之间存在差异,直观上表现为堆芯功率分布的计算值与实测值存在显著误差。为了提高数值模拟精度,本文开展了基于实测数据融合的堆芯物理模型反演优化方法研究:采用人工神经网络算法,通过大量样本训练建立堆芯物理模型与实测数据物理场之间的显式函数关系;基于三维变分算法和实测数据物理场,建立物理模型反演优化代价函数,通过实测数据反演优化得到与真实状态更为接近的堆芯物理模型。为了实现方法验证,本文利用国内某商用压水堆核电厂的功率分布实测数据对堆芯燃料组件弯曲实现了反演优化。数值结果表明:采用反演优化得到的堆芯物理模型,可将堆芯功率分布计算误差的最大值由13.4%降至7.7%,显著提升了堆芯数值模拟结果的精度。因此,本文提出的基于实测数据融合的堆芯物理模型反演优化方法能够显著提高堆芯数值模拟的精度,在核反应堆数字孪生技术研发中具有重要的应用价值。 展开更多
关键词 实测数据融合 模型反演优化 三维变分算法 人工神经网络算法
下载PDF
基于改进MobileNetV2的棉花颜色分级检测 被引量:1
17
作者 王中璞 吴正香 +2 位作者 尤美路 张立杰 阿不都热西提·买买提 《棉纺织技术》 CAS 2024年第6期15-21,共7页
针对棉花颜色级检验中感官检验容易受到主观因素影响、仪器检验不稳定的问题,提出一种使用改进MobileNetV2神经网络实现棉花颜色级检测的方法。通过自主设计的图像采集装置,收集白棉一级到白棉五级5种棉花颜色级样品,制作数据集。将Mobi... 针对棉花颜色级检验中感官检验容易受到主观因素影响、仪器检验不稳定的问题,提出一种使用改进MobileNetV2神经网络实现棉花颜色级检测的方法。通过自主设计的图像采集装置,收集白棉一级到白棉五级5种棉花颜色级样品,制作数据集。将MobileNetV2网络后三层进行特征融合,并嵌入CBAM注意力机制,同时与GhostNet、ShuffleNetV2和原始MobileNetV2模型进行对比,预测棉花颜色分级。结果表明:改进后的MobileNetV2在测试集的准确率达到92.10%,相对于GhostNet、ShuffleNetV2和原始MobileNetV2分别提高了3.01个百分点、4.61个百分点、1.24个百分点,具有较好的检测效果。 展开更多
关键词 MobileNetV2模型 棉花颜色级 神经网络 注意力机制 特征融合
下载PDF
基于高低频特征分解的深度多模态医学图像融合网络 被引量:2
18
作者 王欣雨 刘慧 +2 位作者 朱积成 盛玉瑞 张彩明 《图学学报》 CSCD 北大核心 2024年第1期65-77,共13页
多模态医学图像融合旨在利用跨模态图像的相关性和信息互补性,以增强医学图像在临床应用中的可读性和适用性。然而,现有手工设计的模型无法有效地提取关键目标特征,从而导致融合图像模糊、纹理细节丢失等问题。为此,提出了一种新的基于... 多模态医学图像融合旨在利用跨模态图像的相关性和信息互补性,以增强医学图像在临床应用中的可读性和适用性。然而,现有手工设计的模型无法有效地提取关键目标特征,从而导致融合图像模糊、纹理细节丢失等问题。为此,提出了一种新的基于高低频特征分解的深度多模态医学图像融合网络,将通道注意力和空间注意力机制引入融合过程,在保持全局结构的基础上保留了局部纹理细节信息,实现了更加细致的融合。首先,通过预训练模型VGG-19提取两种模态图像的高频特征,并通过下采样提取其低频特征,形成高低频中间特征图。其次,在特征融合模块嵌入残差注意力网络,依次从通道和空间维度推断注意力图,并将其用来指导输入特征图的自适应特征优化过程。最后,重构模块形成高质量特征表示并输出融合图像。实验结果表明,该算法在Harvard公开数据集和自建腹部数据集峰值信噪比提升8.29%,结构相似性提升85.07%,相关系数提升65.67%,特征互信息提升46.76%,视觉保真度提升80.89%。 展开更多
关键词 多模态医学图像融合 预训练模型 深度学习 高低频特征提取 残差注意力网络
下载PDF
基于模型融合和生成网络的有效阵位智能决策方法
19
作者 郭力强 马亮 +3 位作者 张会 杨静 李连峰 翟雅琪 《系统仿真学报》 CAS CSCD 北大核心 2024年第7期1573-1585,共13页
军事智能技术是当前最具活力的前沿领域和未来无人装备发展的必然趋势。针对无人平台在复杂环境下自主决策可靠性和实时性的双重需求和现有基于规则推演的作战仿真技术在动态性和灵活性方面的不足,采用原理分析与实验验证的研究方法,在... 军事智能技术是当前最具活力的前沿领域和未来无人装备发展的必然趋势。针对无人平台在复杂环境下自主决策可靠性和实时性的双重需求和现有基于规则推演的作战仿真技术在动态性和灵活性方面的不足,采用原理分析与实验验证的研究方法,在某型无人平台射击实验数据集的基础上,围绕攻击决策的有效阵位识别环节,将其转换为机器学习领域类别不平衡的二分类问题,综合采用相关性分析、特征工程、模型融合技术构建高实时性和灵活性的有效阵位智能决策模型,并提出基于ICGAN-Stacking不平衡分类架构对少数类样本进行定向扩充,实现数据增强和模型性能提升。实验结果表明:所提方法召回率提升了4.1%、精确度提升了0.4%、F1值提升了1.5%、AUC值达到90.9%,能够满足无人平台执行作战任务实时性和可靠性需求。 展开更多
关键词 军事智能 无人平台 模型融合 生成对抗网络 不平衡分类
下载PDF
基于听觉特征融合的煤矸识别方法研究
20
作者 杨政 王世博 +4 位作者 饶柱石 杨善国 杨建华 刘送永 刘后广 《振动与冲击》 EI CSCD 北大核心 2024年第8期136-144,共9页
针对强噪声背景下综放开采过程中垮落煤矸难以识别问题,提出了一种融合低级听觉特征Mel频谱和高级听觉特征听觉神经递质发放率的煤矸识别方法。首先,根据煤矸垮落冲击液压支架尾梁声音信号频谱特点,基于听觉神经滤波器组模型构建了适用... 针对强噪声背景下综放开采过程中垮落煤矸难以识别问题,提出了一种融合低级听觉特征Mel频谱和高级听觉特征听觉神经递质发放率的煤矸识别方法。首先,根据煤矸垮落冲击液压支架尾梁声音信号频谱特点,基于听觉神经滤波器组模型构建了适用于煤矸识别任务的听觉模型;然后,利用听觉模型对煤矸垮落声音信号进行分析,获得听觉神经递质发放率;再次,将听觉神经递质发放率与通过Mel频谱提取的峰值特征进行融合,得到煤矸声音听觉感知图;最后,基于所构建的听觉感知图,利用ConvNeXt模型进行煤矸识别。试验结果表明,采用融合听觉特征的煤矸识别方法在不同信噪比下均具有较高的识别准确率;其优越性在背景噪声较大的工况下(信噪比为-5 dB)尤为明显,准确率仍能达到91.52%,显著优于以低级听觉特征和频谱图作为识别特征和利用时频域特征结合机器学习的煤矸识别方法,验证了融合听觉特征的煤矸识别方法对噪声具有优越的鲁棒性。 展开更多
关键词 放顶煤 煤矸识别 听觉模型 听觉神经递质 特征融合 卷积神经网络
下载PDF
上一页 1 2 25 下一页 到第
使用帮助 返回顶部