基于普通克里金外推的迭代维纳滤波法向下延拓

在利用普通克里金方法外推的基础上,通过频域径向平均功率谱估算噪声方差,采用迭代维纳滤波法获取较为稳定的重力场向下延拓结果,并分别利用模拟算例和实际数据进行实验。对于模拟算例,外推后的均方根误差由6.72μGal下降到1.89μGal,相对误差由9.26%下降到4.44%。对于实际数据,外推后延拓结果不仅可避免边缘失真现象,而且能更好地反映浅源高频异常特征信息。

Iterative Wiener Filter Based on Ordinary Kriging Extrapolation for Downward Continuation

Abstract: Based on ordinary Kriging extrapolation method, the noise variance is estimated by frequency radially averaged power spectrum, and then the downward continuation of stable gravity field is calculated by the iterative Wiener filter. The simulation example and actual data are separately used to do experiments. The accuracy of simulation example is significantly improved after extrapolation, the mean square error decreases from 6.72 μGal to 1.89 μGal, and the relative error decreases from 9.26% to 4.44%. The continuation result of actual data after extrapolation reflects the shallower anomaly information with high frequency better and avoids obtaining false values in the boundary.