Adaptive Threshold Shearlet Transform For Microseismic Data Denoising

Microseismic monitoring technology is the geophysical technology that monitors the effects,consequences and underground conditions of production activities by observing and analyzing the generated micro-seismic events in production activities.It is divided into well monitoring and ground monitoring.The noise components in the ground microseismic data are complicated,which makes most of the effective events in phase submerge,which makes it difficult to identify the event signal phase.With the high requirements and high standards for large amounts of ground-based microseismic data,ground-based microseismic exploration technology is constantly being updated.Ground microseismic records with high resolution,high signal-to-noise ratio,and high fidelity are ground-based.Therefore,it is the first problem to be solved in researching and proposing appropriate denoising methods in the process of noise suppression on ground microseismic data,and it has always been the focus of research at home and abroad.The high signal-to-noise ratio is one of the criteria for measuring the quality of terrestrial microseismic data and is the basis of all standards.Therefore,improving the signal-to-noise ratio of microseismic data is the primary task of noise suppression.When the ground microseismic data is processed,if the common seismic exploration denoising method is used to perform microseismic event location and noise suppression on the microseismic data,it will give unsatisfactory results.Although the existing random noise reduction methods have achieved certain results in engineering practice,they generally suppress random noise and also weaken some of the effective signals.Therefore,it is important and difficult in the process of microseismic signal processing to effectively suppress the random noise in terrestrial microseismic data while retaining the effective signal at the maximum.Shearlet transform is a multi-scale,multi-directional and anisotropic decomposition algorithm,which has good processing effect on directional signals and low SNR data.According to the different characteristics of the effective signal and random noise in the transform domain,the Shearlet coefficients can be processed by the threshold to achieve the purpose of denoising.Among them,the threshold has akey role,and using a single threshold to process the coefficients with different characteristics will make the denoised There are either many noise residues in the signal,or serious loss of active ingredients.Therefore,in view of the existing problems and the characteristics of ground microseismic data,this paper proposes a new microseismic data denoising algorithm based on adaptive threshold Shearlet transform.The new algorithm calculates the basic threshold of each subband direction under the SURE criterion,and introduces adjustment factors according to the coefficients of the transform domain and its surrounding neighborhood coefficients,so as to use this factor to adaptively constrain the size of the basic threshold,and finally obtain each transform domain.The self-adaptive threshold of the coefficient realizes the use of different thresholds and threshold processing techniques for the coefficients with different characteristics,and solves the problem that the noise suppression caused by the fixed threshold processing in the conventional method is not ideal.This paper comprehensively analyzes the basic principle of Shearlet transform and the distribution characteristics of effective signals and random noise in simulated and actual terrestrial microseismic records.The advantages and disadvantages of the proposed algorithm in noise suppression are illustrated by the waveforms and spectral comparisons before and after denoising in a single-channel simulation experiment.The proposed algorithm is applied to the noise suppression of analog records and actual terrestrial microseismic records,and compared with the traditional filtering methods.The experimental results show that this method is superior to other methods in suppressing random noise and protecting effective signals,solves the problem that the noise suppression caused by the fixed threshold processing in the conventional method is not ideal.