FIR与PWVD串联滤波消除随机噪声

Using FIR plus PWVD filtering to eliminate random noise.

本文分析了PWVD(Pseudo Wigner-Ville Distribution)消除随机噪声的基本原理,指出其迭代算法计算量大,且在低信噪比资料条件下,局部线性条件难以满足等问题,提出先应用FIR滤波对地震资料进行滤波,使PWVD滤波的应用条件得到满足,再应用PWVD滤波拟取代迭代PWVD滤波算法。这种方法既减少了计算量也改善滤波效果。通过应用雷克子波、理论模型和实际资料对该方法进行试算表明:该方法去噪优于迭代PWVD、FIR滤波和其他多道预测去噪方法。

The paper analyzed the basic mechanism of PWVD eliminating random noise, pointed out the issues of the method having larger computational works of iterative algorithm and being difficult to meet the linear condition locally in a condition of lower S/N ratio of data, presented using FIR filtering to carry out filtering of seismic data firstly, making the application condition of PWVD filtering be met, and then using PWVD filtering to replace iterative PWVD filtering algorithm. The method can both reduce computational works and improve filtering effect. The experimental computation by Ricker wavelet, theoretical model and practical data showed that using the method for denoise is superior to PWVD, FIR filtering and other multi-channel predictive denoise methods.