基于Contourlet系数相关性的地震噪声压制方法

Seismic noise suppression method based on correlation of Contourlet coefficients

地震信号Contourlet变换的稀疏表达中,变换域关联系数具有较强的相关性,据此提出了一种基于Contourlet邻域系数相关性的自适应阈值地震去噪方法。该方法对每一点系数计算其邻域均方根振幅,利用这一点的系数值及其邻域均方根振幅计算其邻域系数相关性,根据邻域系数相关性自适应地确定阈值。数值模拟和实际资料的处理结果验证了本文方法的有效性,该方法在压制噪声的同时能够较好地保护有效信号。

In seismic signal sparse expressions of Contourlet transform, the relevance coefficients in transform domain have strong correlation. An adaptive threshold denoising method based on the correlation of Contourlet neighborhood coefficients is put forward. The neighborhood root-mean-square amplitude coefficients for every point are calculated with the method, the correlation of Contourlet neighborhood coefficients are calculated using the coefficients of the point and neighborhood root-mean-square amplitude and then the threshold according to the correlation of neighborhood coefficients is adaptively determined in this method. Numerical simulation and actual data processing results show the effectiveness of the proposed method, which indicates that it can protect the effective signal in suppressing noise.