Microseismic Signal Reconstruction And Denoising Based On Compressed Sensing

The reserves of unconventional oil and gas resources in China are considerable,but due to the lack of mature exploration and development methods,how to effectively develop and use them has become an important research topic in the field of oil and gas exploration.Microseismic monitoring technology can accurately detect and locate unconventional oil and gas resources such as underground tight gas and shale gas,so it has broad application prospects.However,due to the limitations of the acquisition conditions,the data acquired in microseismic exploration are often irregular and incomplete with low SNR,which has a serious impact on microseismic data processing and interpretation.Therefore,in the actual work,we need to restore and reconstruct the under-sampling seismic data by using a certain reconstruction algorithm.Based on this problem,the recently emergent theory of compressed sensing in signal processing is applied to the reconstruction and denoising of seismic data in this paper,and the reconstruction results are also analyzed.The theory of compressed sensing breaks through the limitation of the traditional Nyquist theorem on the sampling rate of the original signals and uses specific measurement matrices and reconstruction algorithms which can restore and reconstruct the original signal with high accuracy at a lower sampling rate.In this paper,the Curvelet transform is used because its locality and anisotropy are preferable for the sparsity of unsteady signals such as microseismic data.In this paper,the threshold noise attenuation method based on the Curvelet transform,which is seldom used in microseismic monitoring,is introduced into microseismic monitoring,and desirable effect of noise attenuation is obtained.The POCS iterative algorithm is used to successfully restore and reconstruct the conventional seismic signals.Currently,the threshold model used in POCS iterative algorithm is usually exponential.In this paper,the data-driven model,which is often used in image signal processing,is improved and applied to the microseismic signals.Compared with the exponential threshold model,the improved data-driven model has faster speed of convergence and can achieve more desirable reconstruction effect within less time.It is proved by experiments that the improved threshold model is reasonable and has the potential to be applied to the microseismic data.