The 4th International Symposium on Mine Safety Science and Engineering (ISMSSE2018)

305 / 2018-05-03 10:19:03

An algorithm for suppressing noise in seismic data based on compressed sensing

Seismic data; Suppressing noise; Compressive sensing; Block sparse representation; K-singular value decomposition;

Mine Dynamics Disaster > Rock burst impact

The noise introduced in the process of seismic exploration caused serious distortion and interference to the seismic signal, and the conventional seismic data denoising methods can not meet the requirements of high precision seismic exploration. For this reason, an algorithm for suppressing noise in seismic data based on compressed sensing is proposed. According to the local direction feature of seismic data, the seismic data space is divided into multiple subspaces, and the K- singular value decomposition algorithm is used to study the analytical dictionary in each subspace, so as to realize the optimal sparse representation of different subspace data blocks. In the process of seismic data reconstruction, each data block is estimated in all subspaces, and then the optimal estimation of each data block is obtained according to the minimum error criterion of sparse representation, then the seismic data are reconstructed to achieve the purpose of suppressing noise. It is applied to seismic data with different signal-to-noise ratios, and compared with conventional methods of seismic data denoising, the experiment result shows that the proposed method can effectively reduce the noise in seismic data.