子空间聚类在最近几年受到了广泛的关注,新近提出的自适应图卷积子空间聚类方法取得了很好的效果。但是该方法仅适用于单一视图的子空间聚类问题。本文将该方法拓展到多视图上,提出了多视图图卷积子空间聚类。该方法构建了F范数正则项...子空间聚类在最近几年受到了广泛的关注,新近提出的自适应图卷积子空间聚类方法取得了很好的效果。但是该方法仅适用于单一视图的子空间聚类问题。本文将该方法拓展到多视图上,提出了多视图图卷积子空间聚类。该方法构建了F范数正则项以便更有效地挖掘每个视图中数据之间的关系,还构建了不同视图之间的加权机制来融合不同视图之间的信息。大量的实验证明,我们的方法是有效的。Subspace clustering has received extensive attention in recent years. Although the recently proposed adaptive graph convolutional subspace clustering performs well, it can only be applied to subspace clustering problem with a single view. This paper proposes multi-view graph convolutional sub-space clustering to extend this method to the multi-view situation. This method not only constructs F-norm regularization to more effectively mine the relationships between data in each view, but also builds a weighting strategy between different views to integrate their information. A large number of experiments have proved that our method is effective.展开更多
Radon transform is to use the speed difference between primary wave and multiple wave to focus the difference on different"points"or"lines"in Radon domain,so as to suppress multiple wave.However,th...Radon transform is to use the speed difference between primary wave and multiple wave to focus the difference on different"points"or"lines"in Radon domain,so as to suppress multiple wave.However,the limited migration aperture,discrete sampling,and AVO characteristics of seismic data all will weaken the focusing characteristics of Radon transform.In addition,the traditional Radon transform does not take into account the AVO characteristics of seismic data,and uses L1 Norm,the approximate form of L0 Norm,to improve the focusing characteristics of Radon domain,which requires a lot of computation.In this paper,we combine orthogonal polynomials with the parabolic Radon transform(PRT)and find that the AVO characteristics of seismic data can be fitted with orthogonal polynomial coefficients.This allows the problem to be transformed into the frequency domain by Fourier transform and introduces a new variable,lambda,combining frequency and curvature.Through overall sampling of lambda,the PRT operator only needs to be calculated once for each frequency,yielding higher computational efficiency.The sparse solution of PRT under the constraints of the smoothed L0 Norm(SL0)obtained by the steepest descent method and the gradient projection principle.Synthetic and real examples are given to demonstrate that the proposed method has This method has advantages in improving the Radon focusing characteristics than does the PRT based on L1 norm.展开更多
针对三轴磁通门传感器非正交、灵敏度不一致、零偏以及构成梯度仪的两个磁通门传感器位置不一致引起的测量误差问题,建立了误差模型;基于地磁矢量模值短时不变原理,采用线性化最小二乘算法进行一个磁通门传感器参数的辨识;基于校准后三...针对三轴磁通门传感器非正交、灵敏度不一致、零偏以及构成梯度仪的两个磁通门传感器位置不一致引起的测量误差问题,建立了误差模型;基于地磁矢量模值短时不变原理,采用线性化最小二乘算法进行一个磁通门传感器参数的辨识;基于校准后三分量差值F范数最小原理,采用多元线性回归的方法进行另一个磁通门传感器参数以及两个磁通门传感器相对位置关系参数的辨识。实验结果表明,该方法能够将两个磁通门中最大总场误差从1 194.4 n T降低到30.0 n T,将三分量梯度仪最大输出误差从529.1 n T降低到13.4 n T,有效地改善了梯度仪性能。展开更多
文摘子空间聚类在最近几年受到了广泛的关注,新近提出的自适应图卷积子空间聚类方法取得了很好的效果。但是该方法仅适用于单一视图的子空间聚类问题。本文将该方法拓展到多视图上,提出了多视图图卷积子空间聚类。该方法构建了F范数正则项以便更有效地挖掘每个视图中数据之间的关系,还构建了不同视图之间的加权机制来融合不同视图之间的信息。大量的实验证明,我们的方法是有效的。Subspace clustering has received extensive attention in recent years. Although the recently proposed adaptive graph convolutional subspace clustering performs well, it can only be applied to subspace clustering problem with a single view. This paper proposes multi-view graph convolutional sub-space clustering to extend this method to the multi-view situation. This method not only constructs F-norm regularization to more effectively mine the relationships between data in each view, but also builds a weighting strategy between different views to integrate their information. A large number of experiments have proved that our method is effective.
基金funded by the National Natural Science Foundation of China(No.41774133)major national science and technology projects(No.2016ZX05024-003 and 2016ZX05026-002-002)the talent introduction project of China University of Petroleum(East China)(No.20180041)
文摘Radon transform is to use the speed difference between primary wave and multiple wave to focus the difference on different"points"or"lines"in Radon domain,so as to suppress multiple wave.However,the limited migration aperture,discrete sampling,and AVO characteristics of seismic data all will weaken the focusing characteristics of Radon transform.In addition,the traditional Radon transform does not take into account the AVO characteristics of seismic data,and uses L1 Norm,the approximate form of L0 Norm,to improve the focusing characteristics of Radon domain,which requires a lot of computation.In this paper,we combine orthogonal polynomials with the parabolic Radon transform(PRT)and find that the AVO characteristics of seismic data can be fitted with orthogonal polynomial coefficients.This allows the problem to be transformed into the frequency domain by Fourier transform and introduces a new variable,lambda,combining frequency and curvature.Through overall sampling of lambda,the PRT operator only needs to be calculated once for each frequency,yielding higher computational efficiency.The sparse solution of PRT under the constraints of the smoothed L0 Norm(SL0)obtained by the steepest descent method and the gradient projection principle.Synthetic and real examples are given to demonstrate that the proposed method has This method has advantages in improving the Radon focusing characteristics than does the PRT based on L1 norm.
文摘针对三轴磁通门传感器非正交、灵敏度不一致、零偏以及构成梯度仪的两个磁通门传感器位置不一致引起的测量误差问题,建立了误差模型;基于地磁矢量模值短时不变原理,采用线性化最小二乘算法进行一个磁通门传感器参数的辨识;基于校准后三分量差值F范数最小原理,采用多元线性回归的方法进行另一个磁通门传感器参数以及两个磁通门传感器相对位置关系参数的辨识。实验结果表明,该方法能够将两个磁通门中最大总场误差从1 194.4 n T降低到30.0 n T,将三分量梯度仪最大输出误差从529.1 n T降低到13.4 n T,有效地改善了梯度仪性能。
