基于快速加权Schatten P范数最小化的地震数据随机噪声压制

Random Noise Suppression for Seismic Data Based on Fast Weighted Schatten P-norm Minimization

为了有效压制地震数据随机噪声,同时为了提高计算效率,提出了快速加权Schatten P范数最小化(FWSNM)算法。在地震数据非局部自相似块结构的基础上,建立低秩压制噪声模型,该模型利用加权Schatten P范数逼近秩。在模型求解中涉及奇异值分解,利用随机奇异值分解代替奇异值分解,降低算法计算复杂度,以得到快速加权Schatten P范数最小化(FWSNM)算法。实验结果表明,在保持信噪比一致的条件下,FWSNM算法相对于WSNM算法耗时更短。因此FWSNM算法计算效率更高,压制噪声性能更优。

In order to effectively suppress random noise of seismic data,and improve the computational efficiency at the same time,a fast weighted Schatten P-norm minimization(FWSNM)algorithm is proposed in this paper.On the basis of non-locally self-similar block structure of seismic data,a low rank model of noisy suppression is built,and the weighted Schatten P-norm is used to approximate rank in this model.The singular value decomposition is involved in solving the model,and it is replaced by random one,which reduces the computational complexity of the algorithm and improves the computational efficiency.Thus fast weighted Schatten P-norm minimization algorithm is obtained.The experimental results show that FWSNM algorithm takes less time than WSNM algorithm under the condition of keeping the signal to noise ratio consistent.Therefore,FWSNM algorithm has higher computational efficiency and better performance on noisy suppression.