Adaptive Threshold Based On Cycle Spinning Shearlet Transform For Microseismic Random Noise Attenuation

With the constant demand of world oil and gas,oil and gas are consumed as non renewable resources.The development of complex unconventional oil and gas reservoirs has become a hot issue for the growth of reserves.As the observation signal of unconventional hydraulic fracturing,microseismicity has the characteristics of low energy and low signal-to-noise ratio(SNR).Conventional filtering methods have limitations,which seriously affects the accuracy of the first pick-up point and inversion of microseismic signals.It brings difficulties to the interpretation of geological structure and development of oil and gas reservoirs.Therefore,improving the SNR is an important step in microseismic data processing.We need to extract the useful signal from the strong random noise background and effectively reduce the random noise in the microseismic data.In this paper,we focus on the random noise suppression method for microseismic signals.Combined with the Shearlet transform and the cycle spinning transform,we construct an adaptive threshold estimation scheme.The synthetic and field data processing results show that the method can effectively extract the microseismic signal,remove the random noise to a large extent and maintain the effective signal amplitude.Shearlet transform is a new multi-scale,time-frequency analysis tool,which has the characteristics of multi-scale,multi-direction and best sparse approximation.That is,to reconstruct the useful signals by sparse matrix and it has high computational rprocessing and achieved some results.However,Shearlet transform method based on the threshold in the microseismic noise elimination process has certain limitations:firstly,downhole microseismic signal has the characteristics of weak energy and its dominant frequency is relatively high.Effective signal and noise overlap in high frequency band.Threshold based methods are difficult to separate the signal and noise.Secondly,as for the ground microseismic data,the current threshold denoising method based Shearlet transform often assumes that the noise is distributed in the high frequency band,and does not take into account the low-frequency noise interference.Thirdly,the traditional threshold based Shearlet transform denoising method uses uniform threshold in the transform domain,so it is easy to lose some effective signal,resulting in energy loss.Fourthly,the traditional Shearlet transform exist down sampling process,so the transform lack of shift invariance.This paper starts from the essence of Shearlet transform,and studies on microseismic noise suppression technology based on Shearlet transform.The coefficient distribution characteristics of microseismic data after Shearlet transform are analyzed.Considering the direction and spatial correlation of the signal,process the data with cycle spinning,using the multi-scale characteristic of Shearlet transform to enhance its translation invariance.The multi-scale and multi-direction decomposition of the signal is performed and the weighted threshold shrinkage scheme based on the block principal component analysis is established according to the difference of the microseismic signal and the random noise in the Shearlet domain.Finally,the data of the spatial arrangement is superimposed to achieve the purposes of enhancing the effective signal.It can recover the amplitude of the effective signal and meanwhile realize the recognition of the microseismic signal at low SNR.Synthetic and real data process verify the adaptive threshold Shearlet transform algorithm proposed in this paper is better than the threshold based tradional Shearlet transform algorithm in the extent to maintain amplitude and noise suppression.