P-P WAVE AND P-S CONVERTED WAVE SEPARATION AND DE-NOISING COMBINING WITH SVD AND F-K FILTERING

Wave model separation and de-noising of P-P waves and P-S converted waves is a vital task in seismic data processing. In a multi-mode seismic wave field, P-P waves have the dissimilar characteristics in comparison with other seismic signals: the energy is relatively strong, and the propagation law and coherence are more obvious to be observed (with a hyperbolic law). As for P-S converted events, they have large differences in velocity and frequency characteristics compared to other seismic wave modes. This article puts forward a P-P wave, P-S converted wave separation and de-noising workflow by combining Singular Value Decomposition (SVD) and f-k filtering. First, align P-P waves by NMO correction to make they achieve the best coherence in the transverse direction. Then, by extracting the singular values of the P-P waves through SVD, the P-P waves can be separated. Finally, by using an f-k filtering method, P-S converted waves and other noises can be separated. In this way, the complementary advantages of the SVD and the f-k filtering techniques are combined in wave field separation and de-noising, and can avoid the ineffectiveness in separation the P-P waves and P-S converted waves by only SVD or only an f-k filtering technique. Model and actual seismic data have been processed with this method, and good results have been achieved.