基于K-L变换的微地震资料去噪方法

A denoising method based on K-L transform for microseismic data

K-L变换是一种有效的地震去噪方法,但对于信号弱、信噪比较低的微地震资料,传统K-L变换方法去噪效果较差。为此,提出基于K-L变换的微地震资料去噪方法。首先利用单道微地震记录求取协方差矩阵,然后对协方差矩阵进行求解特征值和特征向量,通过分析,选取适当的特征值所对应的新分量进行矩阵的重构,最后通过K-L反变换得到去除噪声、有效信号突出的重构信号。该方法应用于四川某地区微地震资料的去噪处理,效果良好。

The Karhunen -Loeve （K- L） transform is an effective denoising method for seismic data. For microseis- mic data with weak signal and low signal/noise ratio ,however, it is difficult to obtain a satisfactory processing result by u- sing the conventional K - L transform method. Therefore, a microseismic denoising method based on K - L transform was put forward. Firstly, a covariance matrix was built up by using single - channel seismic records. And then eigenvalues and eigenvectors were obtained by solving the covariance matrix. Secondly, new components corresponding to appropriate eigen- values were selected in order to reconstructing a new matrix. Finally, signals with removing noise and highlighting effective signals were reconstructed by using K - L inverse transform. The method has been successfully applied in the denoising process for the microseismic data from a block in Sichuan province.