期刊文献+
任意字段
题名或关键词
题名
关键词
摘要
作者
第一作者
机构
刊名
分类号
参考文献
作者简介
基金资助
栏目信息
共找到2,110篇文章
< 1 2 … 106 >
每页显示 20 50 100
已选0条
引用分析
参考文献
引证文献
 统计分析
检索结果
已选文献
显示方式：
Investigation of the Tikhonov Regularization Method in Regional Gravity Field Modeling by Poisson Wavelets Radial Basis Functions 认领 引用 被引量：3
1
作者 Yihao Wu Bo Zhong Zhicai Luo 《Journal of Earth Science》 SCIE CAS CSCD 2018年第6期1349-1358,共10页
The application of Tikhonov regularization method dealing with the ill-conditioned problems in the regional gravity field modeling by Poisson wavelets is studied. In particular, the choices of the regularization matri... The application of Tikhonov regularization method dealing with the ill-conditioned problems in the regional gravity field modeling by Poisson wavelets is studied. In particular, the choices of the regularization matrices as well as the approaches for estimating the regularization parameters are investigated in details. The numerical results show that the regularized solutions derived from the first-order regularization are better than the ones obtained from zero-order regularization. For cross validation, the optimal regularization parameters are estimated from L-curve, variance component estimation（VCE） and minimum standard deviation（MSTD） approach, respectively, and the results show that the derived regularization parameters from different methods are consistent with each other. Together with the firstorder Tikhonov regularization and VCE method, the optimal network of Poisson wavelets is derived, based on which the local gravimetric geoid is computed. The accuracy of the corresponding gravimetric geoid reaches 1.1 cm in Netherlands, which validates the reliability of using Tikhonov regularization method in tackling the ill-conditioned problem for regional gravity field modeling. 展开更多
关键词 regional gravity field modeling Poisson wavelets radial basis functions Tikhonov regularization method L-curve variance component estimation（VCE）
暂未订购 下载PDF
Phase spectrum estimation of the seismic wavelet based on a criterion function 认领 引用 被引量：4
2
作者 Yu Yongcai Wang Shangxu +1 位作者 Yuan Sanyi Qi Pengfei 《Petroleum Science》 SCIE CAS CSCD 2012年第2期170-181,共12页
Phase spectrum estimation of the seismic wavelet is an important issue in high-resolution seismic data processing and interpretation. On the basis of two patterns of constant-phase rotation and root transform for wave... Phase spectrum estimation of the seismic wavelet is an important issue in high-resolution seismic data processing and interpretation. On the basis of two patterns of constant-phase rotation and root transform for wavelet phase spectrum variation, we introduce six sparse criteria, including Lu’s improved kurtosis criterion, the parsimony criterion, exponential transform criterion, Sech criterion, Cauchy criterion, and the modified Cauchy criterion, to phase spectrum estimation of the seismic wavelet, obtaining an equivalent effect to the kurtosis criterion. Through numerical experiments, we find that when the reflectivity is not a sparse sequence, the estimated phase spectrum of the seismic wavelet based on the criterion function will deviate from the true value. In order to eliminate the influence of non-sparse reflectivity series in a single trace, we apply the method to the multi-trace seismogram, improving the accuracy of seismic wavelet phase spectrum estimation. 展开更多
关键词 Seismic wavelet phase spectrum criterion function sparse
暂未订购 下载PDF
UNIFORM ANALYTIC CONSTRUCTION OF WAVELET ANALYSIS FILTERS BASED ON SINE AND COSINE TRIGONOMETRIC FUNCTIONS 认领 引用 被引量：1
3
作者 李建平 唐远炎 +1 位作者 严中洪 张万萍 《Applied Mathematics and Mechanics(English Edition)》 SCIE EI 2001年第5期569-585,共17页
Based on sine and cosine functions, the compactly supported orthogonal wavelet filter coefficients with arbitrary length are constructed for the first time. When N = 2(k-1) and N = 2k, the unified analytic constructio... Based on sine and cosine functions, the compactly supported orthogonal wavelet filter coefficients with arbitrary length are constructed for the first time. When N = 2(k-1) and N = 2k, the unified analytic constructions of orthogonal wavelet filters are put forward, respectively. The famous Daubechies filter and some other well-known wavelet filters are tested by the proposed novel method which is very useful for wavelet theory research and many application areas such as pattern recognition. 展开更多
关键词 wavelet analysis filter trigonometric functions analytic construction
暂未订购 下载PDF
Inversion of receiver function by wavelet transformation 认领 引用 被引量：1
4
作者 吴庆举 田小波 +2 位作者 张乃铃 李桂银 曾融生 《Acta Seismologica Sinica(English Edition)》 2003年第6期616-623,共8页
A new method for receiver function inversion by wavelet transformation is presented in this paper. Receiver func-tion is expanded to different scales with different resolution by wavelet transformation. After an initi... A new method for receiver function inversion by wavelet transformation is presented in this paper. Receiver func-tion is expanded to different scales with different resolution by wavelet transformation. After an initial model be-ing taken, a generalized least-squares inversion procedure is gradually carried out for receiver function from low to high scale, with the inversion result for low order receiver function as the initial model for high order. A neighborhood containing the global minimum is firstly searched from low scale receiver function, and will gradu-ally focus at the global minimum by introducing high scale information of receiver function. With the gradual ad-dition of high wave-number to smooth background velocity structure, wavelet transformation can keep the inver-sion result converge to the global minimum, reduce to certain extent the dependence of inversion result on the initial model, overcome the nonuniqueness of generalized least-squares inversion, and obtain reliable crustal and upper mantle velocity with high resolution. 展开更多
关键词 receiver function wavelet transformation waveform inversion
暂未订购 下载PDF
APPLICATION OF WAVELET THEORY IN RESEARCHON WEIGHT FUNCTION OF MESHLESS METHOD 认领 引用 被引量：1
5
作者 张红 张选兵 葛修润 《Applied Mathematics and Mechanics(English Edition)》 SCIE EI 2005年第5期662-666,共5页
Multiresolution analysis of wavelet theory can give an effective way to describe the information at various levels of approximations or different resolutions, based on spline wavelet analysis,so weight function is ort... Multiresolution analysis of wavelet theory can give an effective way to describe the information at various levels of approximations or different resolutions, based on spline wavelet analysis,so weight function is orthonormally projected onto a sequence of closed spline subspaces, and is viewed at various levels of approximations or different resolutions. Now, the useful new way to research weight function is found, and the numerical result is given. 展开更多
关键词 meshless method weight function spline wavelet multiresolution analysis
暂未订购 下载PDF
Cylindrical Guided Wave Signals for Underground Pipe Inspection using Different Continuous Wavelet Mother Functions 认领 引用 被引量：1
6
作者 Rais Ahmad Tribikram Kundu 《Journal of Civil Engineering and Architecture》 2011年第12期1103-1110,共8页
Over the past decade, wavelets provided a powerful and flexible set of tools for handling fundamental problems in science and engineering. Wavelet analyses are being used for solving problems in different engineering ... Over the past decade, wavelets provided a powerful and flexible set of tools for handling fundamental problems in science and engineering. Wavelet analyses are being used for solving problems in different engineering areas like audio de-noising, signal compression, object detection, image decomposition, speech recognition etc. Wavelet analysis employs orthonormal as well as non-orthonornal functions. This research investigates the effectiveness of wavelet analysis in detecting defects in underground steel pipe networks. Continuous Wavelet Transforms （CWT） has been performed on the received signals of cylindrical guided waves. Cylindrical Guided waves are generated and propagated through the pipe wall boundaries in a pitch-catch system. Piezo-electric transducers are used to generate as well as receive guided waves. Several mother wavelet functions such as Daubechies, Symlet, Coiflet and Meyer have been used for the Continuous Wavelet Transform to investigate the most suitable function for defect detection. This research also investigates the effect of surrounding soil on wavelet transforms for different mother wavelet functions. 展开更多
关键词 Wavelet analysis guided waves mother wavelet functions continuous wavelet transform.
暂未订购 下载PDF
THE WAVELET TRANSFORM OF PERIODIC FUNCTION AND NONSTATIONARY PERIODIC FUNCTION 认领 引用
7
作者 刘海峰 周炜星 +2 位作者 王辅臣 龚欣 于遵宏 《Applied Mathematics and Mechanics(English Edition)》 SCIE EI 2002年第9期1062-1070,共9页
Some properties of the wavelet transform of trigonometric Junction, periodic function and nonstationary periodic function have been investigated. The results show that the peak height and width in wavelet energy spect... Some properties of the wavelet transform of trigonometric Junction, periodic function and nonstationary periodic function have been investigated. The results show that the peak height and width in wavelet energy spectrum of a periodic function are in proportion to its period. At the same time, a new equation, which can truly reconstruct a trigonometric function with only one scale wavelet coefficient, is presented. The reconstructed wave shape of a periodic function with the equation is better than any term of its Fourier series. And the reconstructed wave shape of a class of nonstationary periodic function with this equation agrees well with the function. 展开更多
关键词 wavelet transform periodic function nonstationary periodic function Fourier transform
暂未订购 下载PDF
CHARACTERIZATIONS OF FUNDAMENTAL SCALING FUNCTIONS AND WAVELETS 认领 引用
8
作者 Charles K.Chui 《Analysis in Theory and Applications》 1993年第3期37-52,共16页
The objective of Ibis paper is to establish precise characterizations of scaling functions which are orthonormal or fundamental.A criterion for the corresponding wavelets is also given.
关键词 CHARACTERIZATIONS OF FUNDAMENTAL SCALING FUNCTIONS AND WAVELETS Mallat
暂未订购 下载PDF
Local Extrema of Periodic Function’s Wavelet Transform 认领 引用 被引量：3
9
作者 FAN Qi-bin SONG Xiao-yan 《Wuhan University Journal of Natural Sciences》 EI CAS 2005年第6期949-952,共4页
The theory of detecling ridges in the modulus of the continuous wavelet transform is presented as well as reconstructing signal by using information on ridges,To periodic signal we suppose Morlet wavelet as basic wave... The theory of detecling ridges in the modulus of the continuous wavelet transform is presented as well as reconstructing signal by using information on ridges,To periodic signal we suppose Morlet wavelet as basic wavelet, and research the local extreme point and extrema of the wavelet transform on periodic function for the collection of signal＇ s instantaneous amplitude and period. 展开更多
关键词 extrema periodic function wavelet transform time-frequency analysis
暂未订购 下载PDF
Detection of Rice Bacterial Leaf Blight Using Hyperspectral Technology and Continuous Wavelet Analysis 认领 引用
10
作者 Kaihao Shi Lin Yuan +5 位作者 Qimeng Yu Zhongting Shen Yingtan Yu Chenwei Nie Xingjian Zhou Jingcheng Zhang 《Phyton-International Journal of Experimental Botany》 SCIE 2025年第7期2033-2054,共22页
Plant diseases are a major threat that can severely impact the production of agriculture and forestry.This can lead to the disruption of ecosystem functions and health.With its ability to capture continuous narrow-ban... Plant diseases are a major threat that can severely impact the production of agriculture and forestry.This can lead to the disruption of ecosystem functions and health.With its ability to capture continuous narrow-band spectra,hyperspectral technology has become a crucial tool to monitor crop diseases using remote sensing.However,existing continuous wavelet analysis(CWA)methods suffer from feature redundancy issues,while the continuous wavelet projection algorithm(CWPA),an optimization approach for feature selection,has not been fully validated to monitor plant diseases.This study utilized rice bacterial leaf blight(BLB)as an example by evaluating the performance of four wavelet basis functions-Gaussian2,Mexican hat,Meyer,andMorlet-within theCWAandCWPAframeworks.Additionally,the classification models were constructed using the k-nearest neighbors(KNN),randomforest(RF),and Naïve Bayes(NB)algorithms.The results showed the following:(1)Compared to traditional CWA,CWPA significantly reduced the number of required features.Under the CWPA framework,almost all the model combinations achieved maximum classification accuracy with only one feature.In contrast,the CWA framework required three to seven features.(2)Thechoice of wavelet basis functions markedly affected the performance of themodel.Of the four functions tested,the Meyer wavelet demonstrated the best overall performance in both the CWPA and CWA frameworks.(3)Under theCWPAframework,theMeyer-KNNandMeyer-NBcombinations achieved the highest overall accuracy of 93.75%using just one feature.In contrast,under the CWA framework,the CWA-RF combination achieved comparable accuracy(93.75%)but required six features.This study verified the technical advantages of CWPA for monitoring crop diseases,identified an optimal wavelet basis function selection scheme,and provided reliable technical support to precisely monitor BLB in rice(Oryza sativa).Moreover,the proposed methodological framework offers a scalable approach for the early diagnosis and assessment of plant stress,which can contribute to improved accuracy and timeliness when plant stress is monitored. 展开更多
关键词 Hyperspectral continuous wavelet analysis continuous wavelet projection algorithm wavelet basis function disease monitoring
暂未订购 下载PDF
Wavelet-Based Fractal Function Approximation 认领 引用 被引量：1
11
作者 Zhang Hejei Tao Ran Zhou Siyong & Wang Yue 《Journal of Systems Engineering and Electronics》 EI 1999年第4期60-66,共7页
In this paper, we study on the application of radical B-spline wavelet scaling function in fractal function approximation system. The paper proposes a wavelet-based fractal function approximation algorithm in which th... In this paper, we study on the application of radical B-spline wavelet scaling function in fractal function approximation system. The paper proposes a wavelet-based fractal function approximation algorithm in which the coefficients can be determined by solving a convex quadraticprogramming problem. And the experiment result shows that the approximation error of this algorithm is smaller than that of the polynomial-based fractal function approximation. This newalgorithm exploits the consistency between fractal and scaling function in multi-scale and multiresolution, has a better approximation effect and high potential in data compression, especially inimage compression. 展开更多
关键词 B-spline Wavelet scaling function Fractal function Approximation Quadratic programming.
暂未订购 下载PDF
Gaussian Functions,Γ-Functions and Wavelets 认领 引用 被引量：1
12
作者 蔡涛 许天周 《Journal of Beijing Institute of Technology》 EI CAS 2003年第2期198-201,共4页
The relations between Gaussian function and Γ function is revealed first at one dimensional situation. Then, the Fourier transformation of n dimensional Gaussian function is deduced by a lemma. Following the train of... The relations between Gaussian function and Γ function is revealed first at one dimensional situation. Then, the Fourier transformation of n dimensional Gaussian function is deduced by a lemma. Following the train of thought in one dimensional situation, the relation between n dimensional Gaussian function and Γ function is given. By these, the possibility of arbitrary derivative of an n dimensional Gaussian function being a mother wavelet is indicated. The result will take some enlightening role in exploring the internal relations between Gaussian function and Γ function as well as in finding high dimensional mother wavelets. 展开更多
关键词 Gaussian function Γfunction wavelet
暂未订购 下载PDF
基于可解释模型的低速重载轴承故障诊断 认领 引用 被引量：2
13
作者 孙艳玲 孙显彬 +2 位作者 贾新月 宋益民 于春雨 《轴承》 北大核心 2026年第1期84-90,共7页
针对低速重载轴承低转速导致故障信号微弱,故障特征提取困难的技术难点,以及深度学习由于自身“黑盒”特性导致诊断结果的不可解释和不可信任的问题,构建了一种基于注意力机制和自适应激活函数的小波内核可解释网络模型,以实现低速重载... 针对低速重载轴承低转速导致故障信号微弱,故障特征提取困难的技术难点,以及深度学习由于自身“黑盒”特性导致诊断结果的不可解释和不可信任的问题,构建了一种基于注意力机制和自适应激活函数的小波内核可解释网络模型,以实现低速重载轴承的故障诊断。设计了一个能够自动调整参数的自适应激活函数适应不同的任务,以Morlet小波和Laplace小波内核代替随机卷积核使模型具有理论上的可解释性,引入注意力机制和自适应激活函数提高网络的特征表达能力。通过振动数据与声发射数据驱动可解释网络模型的对比试验表明:可解释网络模型在低速重载轴承故障诊断领域具有诊断精度高、可信任性强等特点;与振动信号相比,基于声发射信号的低速重载轴承故障诊断更具优势。 展开更多
关键词 滚动轴承 深度学习 小波变换 激活函数 故障诊断
暂未订购 下载PDF
基于奇异值分解的复杂边界条件下开口板弯曲振动特性分析 认领 引用
14
作者 杨舟 游春华 +3 位作者 王鹏韬 郭范波 贺辉 夏宇 《振动工程学报》 EI CSCD 北大核心 2026年第7期1922-1929,共8页
以能量法为基本框架,本文提出一种采用奇异值分解处理边界约束条件的开口板弯曲振动特性求解方法,其基本思想是从求解结构边界约束方程角度出发,通过矩阵奇异值分解得到边界约束方程的基础解系,再通过线性组合的方式将基础解系转化为位... 以能量法为基本框架,本文提出一种采用奇异值分解处理边界约束条件的开口板弯曲振动特性求解方法,其基本思想是从求解结构边界约束方程角度出发,通过矩阵奇异值分解得到边界约束方程的基础解系,再通过线性组合的方式将基础解系转化为位移函数的未知系数列向量,进而将有约束问题转化为无约束问题。与人工弹簧法相比,本文方法具有更好的收敛性。此外,选取高斯小波函数作为位移函数,其本身的局域化特性能够准确捕捉到局部开口处的特征。对复杂边界条件下单开口板和多开口板进行分析,结合有限元方法的计算结果,验证了本文方法的准确性和适用性。 展开更多
关键词 开口矩形板 奇异值分解 能量法 高斯小波函数 弯曲振动
暂未订购 下载PDF
耦合断层错动与地震动的隧道抗震输入方法 认领 引用
15
作者 郑兴群 赵伯明 《中国铁道科学》 EI CAS CSCD 北大核心 2026年第1期160-170,共11页
针对现有隧道抗震相关研究大多将形影相随的断层错动和地震动分开考虑的局限性,提出一种随机合成耦合断层错动和地震动的强震动合成方法。首先,借助数值格林函数法计算得出断层错动引起的地表永久位移,参照M&P速度脉冲波形推导得到... 针对现有隧道抗震相关研究大多将形影相随的断层错动和地震动分开考虑的局限性,提出一种随机合成耦合断层错动和地震动的强震动合成方法。首先,借助数值格林函数法计算得出断层错动引起的地表永久位移,参照M&P速度脉冲波形推导得到表征永久位移的小波函数;然后,以小波函数动态化永久位移,并将其加入合成得到的强震记录的位移时程中,实现断层错动与地震动作用耦合。采用三维有限元模型,模拟断层错动单独作用、地震动单独作用、断层错动与地震动共同作用下的隧道结构的地震响应。结果表明:无论何种加载方式,隧道的破坏都主要集中在穿越断层处;隧道结构在地震荷载作用下的受压损伤程度普遍高于受拉损伤,算例中压缩损伤因子超过0.8;单独施加任一荷载都会低估地震作用对隧道结构的破坏作用,耦合荷载作用下的隧道变形显著超越单独荷载作用下的,算例中耦合荷载作用下隧道拱顶永久位移量级达单一地震动工况的2.1倍。 展开更多
关键词 隧道工程 断层错动 地震动 小波函数 有限元模型
暂未订购 下载PDF
EEMD与改进小波阈值联合的齿轮声发射信号降噪 认领 引用
16
作者 魏娜莎 丁泽鹏 +2 位作者 张瑞亮 刘江锋 田志毅 《噪声与振动控制》 CSCD 北大核心 2026年第4期142-148,169,共7页
针对齿轮运行的声发射信号,提出一种将集合经验模态分解(Ensemble Empirical Modal Decomposition,EEMD)与改进小波阈值相结合的降噪方法。首先,基于Sigmoid函数平滑、易于求导的特点,改进传统的小波阈值函数。其次,构建仿真加噪信号进... 针对齿轮运行的声发射信号,提出一种将集合经验模态分解(Ensemble Empirical Modal Decomposition,EEMD)与改进小波阈值相结合的降噪方法。首先,基于Sigmoid函数平滑、易于求导的特点,改进传统的小波阈值函数。其次,构建仿真加噪信号进行EEMD分解,得到一系列本征模函数(Intrinsic Mode Function,IMF)分量,计算各IMF分量的相关性系数并重构信号。最后,采用河马优化(Hippopotamus Optimizatio,HO)算法在改进小波阈值函数中寻找最佳调整参数α,使用改进小波阈值降噪得到最终的去噪信号,以信噪比(Signal-to-Noise Ratio,SNR)、均方误差(Mean Squared Error,MSE)和余弦相似度(Cosine Similarity,Cos)为评价标准,仿真结果显示,所提方法与传统降噪方法相比,信噪比显著提高。在此基础上,用齿轮试验所采集的声发射信号进行验证,试验结果表明:该方法相比于传统的降噪方法有更好的降噪效果,实用价值更高。 展开更多
关键词 声学 声发射 集合经验模态分解 小波阈值函数 Sigmoid函数 河马优化算法
暂未订购 下载PDF
一种新型小波自适应阈值函数及其降噪应用 认领 引用
17
作者 刘明轩 汤赫男 +1 位作者 徐方素 罗园庆 《机械设计》 CSCD 北大核心 2026年第1期100-107,共8页
为了使经典的小波阈值降噪算法更好地兼顾噪声抑制和细节信息保留,提出了一种新型小波阈值函数。以避免函数不连续和减小小波系数恒定偏差为原则,基于双曲正切函数构造了一个具有高阶可导性质和小波系数自适应收缩能力的阈值函数,并应... 为了使经典的小波阈值降噪算法更好地兼顾噪声抑制和细节信息保留,提出了一种新型小波阈值函数。以避免函数不连续和减小小波系数恒定偏差为原则,基于双曲正切函数构造了一个具有高阶可导性质和小波系数自适应收缩能力的阈值函数,并应用粒子群算法在降噪中对其相关参数进行自动寻优。以通常受到强背景噪声影响的轴承故障信息的工程应用为例,采用不同阈值函数对轴承内圈的故障信号进行降噪与试验对比验证。仿真和试验结果表明:构造的阈值函数可实现小波系数自适应收缩,在增强算法降噪能力的同时,提高了算法对混叠在随机噪声中数据细节信息的保真能力,具备较高的工程应用价值。 展开更多
关键词 小波降噪 阈值函数 粒子群优化 故障诊断 信号处理
暂未订购 下载PDF
基于自适应改进小波阈值限幅滤波的平抑风电波动策略 认领 引用
18
作者 李培强 刘抒睿 欧阳旭 《电气工程学报》 CSCD 北大核心 2026年第2期358-370,共13页
基于风电功率的随机性与波动性对电网运行稳定性的负面影响,提出了一种加入限幅环节的改进小波阈值函数滤波,并通过改进的麻雀搜索算法对改进小波阈值函数的调节系数进行不同工业场景下的寻优,得到适应不同需求的波动平抑后并网功率曲... 基于风电功率的随机性与波动性对电网运行稳定性的负面影响,提出了一种加入限幅环节的改进小波阈值函数滤波,并通过改进的麻雀搜索算法对改进小波阈值函数的调节系数进行不同工业场景下的寻优,得到适应不同需求的波动平抑后并网功率曲线。对风电功率进行平抑后,以储能出力积累与波动平抑欧氏距离指标对不同场景下的并网功率曲线进行评估。算例研究表明,所提改进小波阈值函数限幅滤波能够很好平抑风电波动,并使并网功率限制在我国1 min、10 min尺度的MPFR限值以下,且可以适应不同工业需求调节目标函数权重值得到相应的并网功率曲线;同时验证了NTSSA算法在寻优性能上的优越性。 展开更多
关键词 风电功率 小波阈值滤波 阈值函数 波动平抑
暂未订购 下载PDF
基于小波浅层残差网络的SAR图像船舰小目标检测方法 认领 引用
19
作者 陈昕 张晓琪 +4 位作者 廖庭宇 王学伟 唐顺杰 吴梓杰 单慧琳 《航天返回与遥感》 CSCD 北大核心 2026年第3期506-518,共13页
在合成孔径雷达(Synthetic Aperture Radar,SAR)图像船舰小目标检测任务中,由噪声、杂波干扰等因素造成的细节信息丢失和边缘模糊会导致检测精度下降。针对此问题,文章提出了一种基于小波浅层残差网络(Wavelet Shallow Skip Residual Ne... 在合成孔径雷达(Synthetic Aperture Radar,SAR)图像船舰小目标检测任务中,由噪声、杂波干扰等因素造成的细节信息丢失和边缘模糊会导致检测精度下降。针对此问题,文章提出了一种基于小波浅层残差网络(Wavelet Shallow Skip Residual Network,WSSRNet)的SAR图像船舰小目标检测方法。首先,在主干特征提取网络中构建多尺度小波卷积残差模块(Multi-scale Wavelet Transform Residual Module,MWTRM),引入小波变换卷积获取丰富的感受野并融合多尺度特征,以增强对上下文以及空间特征信息的关注,在降低背景干扰的同时减少计算开销。其次,提出浅层跳跃残差模块(Shallow Skip Residual Module,SSRM),通过设计浅层特征金字塔结构与跨阶段特征融合,增强对位置信息及浅层细节特征的利用,提升小目标的定位精度与召回率。最后,提出了一种新的损失函数InnerShapeIoU,根据目标的尺度差异自适应调整回归策略,提高网络的泛化性能以及对目标的定位精度。在HRSID数据集上的实验结果表明,文章方法准确率达92.3%,召回率达87.1%,平均检测精度(mean Average Precision,mAP)在IoU阈值为0.5的条件下达94.7%,在0.5~0.95的条件下达70.2%。文章方法在检测精度、参数量及对小目标特征提取能力上表现卓越,多场景检测性能强、鲁棒性佳,能显著改善小目标在复杂场景下的错检、漏检问题。 展开更多
关键词 深度学习 跳跃连接 小尺寸目标检测 Inner-ShapeIoU损失函数 小波变换卷积
暂未订购 下载PDF
基于YOLOX改进模型的金属表面缺陷检测 认领 引用
20
作者 车国霖 傅家辉 《吉林大学学报（理学版）》 CAS 北大核心 2026年第3期603-616,共14页
针对金属表面缺陷检测中模型精度与推理速度难以兼顾的问题,提出一种基于YOLOX模型改进的SWE-YOLOX检测算法.首先,为解决复杂背景干扰大及缺陷尺度变化显著的问题,引入混洗通道注意力模块,增强特征表达能力并抑制无关信息;其次,针对缺... 针对金属表面缺陷检测中模型精度与推理速度难以兼顾的问题,提出一种基于YOLOX模型改进的SWE-YOLOX检测算法.首先,为解决复杂背景干扰大及缺陷尺度变化显著的问题,引入混洗通道注意力模块,增强特征表达能力并抑制无关信息;其次,针对缺陷边缘模糊、纹理不清晰的问题,融合小波卷积以提升频域特征提取能力,从而强化细节信息表达;最后,将原有交并比(IoU)损失函数替换为EIoU损失函数,以优化预测框与真实框的回归精度.实验结果表明,该方法在数据集NEU-DET上平均表面精度(mAP)达76.3%,较YOLOX模型提升3.86百分点,且在参数量与计算复杂度基本不增加的前提下保持了较快的推理速度. 展开更多
关键词 YOLOX模型 注意力模块 小波卷积 损失函数 金属表面缺陷
暂未订购 下载PDF
已选0条
引用分析
参考文献
引证文献
 统计分析
检索结果
已选文献
上一页 1 2 … 106 下一页 到第 页
在线咨询 使用帮助 返回顶部 意见反馈