Random noise suppression of seismic data based on CEEMD-MSSA

Random noise is one of the common noises in seismic data, which has a direct impact on high-resolution imaging processing and fine interpretation of seismic data. Random noise attenuation methods based on the low-rank hypothesis of seismic data have been widely used in noise suppression. However, due to the complexity of seismic data, its suppression effect is difficult to meet the practical needs. To solve the above problems, a random noise suppression technique combining complementary empirical mode decomposition (CEEMD) and multi-channel singular spectrum analysis (MSSA) algorithm is proposed. The technique firstly extracts horizontal signal components from noise-containing seismic data in f-x domain based on CEEMD algorithm, and then extracts oblique signal components based on MSSA algorithm. Finally, the random noise suppression of seismic data is realized by the superposition of horizontal component and inclined component. Based on the low-rank nature of seismic signals, the proposed method makes full use of the advantages of CEEMD and MSSA algorithms in horizontal and oblique signal component recognition, which is conducive to the effective detection and extraction of seismic signals, so as to improve the signal-to-noise ratio. The synthetic and practical data application show that compared with the traditional MSSA and EMD-MSSA algorithms, the proposed method has a better effect on random noise suppression, and the signal-to-noise ratio of seismic data is significantly improved. It can provide high-quality data input for subsequent seismic data processing, and has important practical application value.