地震随机噪声衰减的改进经验模态分解法研究及应用

Research and application of improved empirical mode decomposition method for seismic random noise attenuation

为解决地震随机噪声衰减问题,引入经验模态分解法,但该方法存在严重的端点效应问题并使得结果存在较大的误差。因此提出通过寻找包络线极大值点并进行平行延拓从而修正边界极值点的方法来对经验模态分解法进行改进。通过仿真实验和理论模型的测试,发现改进后的方法能很好地抑制端点效应,分解更精准,在对地震随机噪声进行衰减同时能更好地保护构造信息。

In order to solve the problem of seismic random noise attenuation,the empirical mode decomposition method is introduced,but the method has serious endpoint effects and the results have large errors.Therefore,the empirical mode decomposition method is improved by finding the maximum point of the envelope line and conducting parallel extension to correct the boundary extreme point.Through the simulation experiment and theoretical model test,it is found that the improved method can well suppress the endpoint effect,decompose more accurately,and protect the structural information while attenuating the seismic random noise.