Application Of Low-rank Algorithm Based On Transform Domain In Desert Seismic Signal Analysis

High-quality seismic data is the basis of Stratigraphic imaging and interpretation,but the random noise will greatly affect the quality of seismic data.As the signal-to-noise ratio of the seismic data decreases,the characteristics of the noise become more complicated.Especially in desert area,the desert area is rich in oil and gas resources.However,the low signal-to-noise ratio of seismic data with the complicated and unpredictable random noise have brought great difficulties on geophysical structural imaging and interpretation.Therefore,noise reduction is the primary task of desert seismic exploration.The random noise in desert seismic records is mainly concentrated in the low frequency band,which causes serious overlap of noise and effective signal in frequency and amplitude.In addition,the desert noise does not completely obey the Gaussian distribution,so the most methods for Gaussian noise suppression are not suitable for the suppression of desert noise.Therefore,the suppression of desert noise is extremely challenging.In order to improve the signal-to-noise ratio,experts of signal processing from various countries around the world are constantly researching in recent years.A variety of methods have been proposed and effectively applied on signal extraction and noise suppression of seismic data.The representative of these methods are mostly the methods of multi-scale analysis and time-frequency analysis.Based on the above,this paper proposes a low-rank denoising method based on transform domain.The algorithm uses synchroaqueezing wavelet transform to transform seismic data to sparse subspaces one by one trace,and then uses low rank decomposition to decompose the low rank part in the sparse subspace.Different from traditional low-rank algorithm,this paper performs an adaptive iterative convergence on the low-rank decomposition algorithm.When the decomposition error reaches apredetermined range,the effective low-rank component is extracted.Finally,the signal is converted back to time domain by the inverse synchrosqueezing wavelet transform to achieve the denoising.The processing results of synthetic and actual records verify the effectiveness of the proposed method,and the method can be applied to the denoising of desert seismic signal.In addition,it has a significant suppression effect on surface waves.The advantages of the algorithm can be seen in the comparative experiments.