Research On Denoising Method Of Seismic Signal Based On Modal Decomposition Technology

Exploration stratigraphic structure need to research and analyze seismic signals,however,are usually collected seismic signals not only contains useful information,also there are a large number of different forms of noise,especially the depth of exploration in recent years,more and more big,of geology in the exploration areas are more complex,noise and useful information are intertwined,and serious interference with the analysis and processing,so the seismic signal denoising is particularly important.In this paper,a series of researches on the removal of random noise from seismic signals are carried out based on modal decomposition technology.The main research contents are as follows:First of all,the basic principles of empirical mode decomposition EMD,set empirical mode decomposition EEMD and complete set empirical mode decomposition CEEMD and the disadvantages of the algorithms themselves are deeply studied.EMD has the problem of modal aliasing and endpoint effect,the original signal of EEMD auxiliary noise pollution and CEEMD residual noise are transmitted from high order to low order.Three kinds of decomposition algorithms are used to decompose the pure signal and the traditional denoising of the simulated seismic signal.Secondly,based on the above mentioned,a new decomposition algorithm,adaptive complete set empirical mode decomposition CEEMDAN,is introduced,and its improved method and decomposition effect are analyzed.It is found that its decomposition effect is better than EMD,EEMD and CEEMD.Insight into the sample entropy and energy entropy concept,threshold denoising method and random noise and useful information is different from the actual signal correlation problems,and puts forward CEEMDAN adaptive threshold denoising method,the actual signal via CEEMDAN decomposition of the IMF component using the sample entropy,high frequency,correlation analysis was carried out on the high frequency component,remove the coefficient of useful information on position,keep random noise coefficient and structure of high frequency component of the denoising threshold,then for the high frequency component after denoising and retain the low-frequency component refactoring to achieve denoising,By comparing with the traditional denoising methods,this method not only solves the problem that the high frequency components of the traditional denoising methods need to be divided artificially,but also has advantages in improving the signal-to-noise ratio and retaining useful information.Finally,the above CEEMDAN adaptive threshold denoising method is found to be less efficient in de-noising actual seismic signals,because CEEMDAN is a recursive iterative decomposition algorithm,and a large number of iterations and aggregate average calculations are needed to decompose seismic signals.The VMD of the adaptive decomposition in the frequency domain was studied and analyzed,and it was found that the decomposition process did not involve a large number of iterations and aggregate average calculations,and the decomposition efficiency was higher.Through the comparative analysis of experimental simulation and CEEMDAN,it was verified to be better in terms of decomposition accuracy,decomposition effect and noise resistance,etc.Thus VMD joint energy entropy threshold denoising method,and by using the denoising method with CEEMDAN adaptive threshold denoising method of the theory of seismic signals and real seismic signal and noise signal containing different SNR noise denoising and compare,verify the method is not only to improve the signal-to-noise ratio and retain useful information is relative to the CEEMDAN adaptive threshold denoising method is superior,and more suitable for low signal-to-noise ratio seismic signal processing,has been greatly narrowed the operation time,improve efficiency.