Reconstruction And Weak Signal Enhancement Technology Of Deep Seismic Data Based On Compressed Sensing

With the gradual expansion of China's exploration area,the goal of oil and gas exploration has gradually turned to complex structures,stratigraphic and lithologic trap reservoirs,and the target layer of exploration has shifted from the shallow or middle to the deep or ultra-deep layers.Thus,the requirements for seismic data processing technology are getting higher and higher.The deep geological conditions are complex,the deep effective signal energy of seismic data is weak,and the spatial aliasing phenomenon is serious.By improving the regularity,signal-to-noise ratio,resolution and fidelity of seismic data,it can provide reliable data for subsequent imaging,full waveform inversion and seismic data interpretation,which is helpful for judging the target reservoir.In recent years,the theory of compressed sensing has broken the traditional Nyquist sampling concept,and utilizes the sparseness or compressibility of signals to achieve accurate signal reconstruction by solving the regularized inversion problem.Since the theory was put forward,some geophysicists have conducted extensive research on it in the field of oil and gas exploration.Complex geological structures lead to complex features and relatively weak effective signals on deep reflection seismic profiles,while traditional reconstruction and denoising methods are difficult to obtain high-quality effective signals.Based on the theory of compressed sensing,the paper develops seismic data reconstruction and weak signal extraction and enhancement algorithms for complex structures.The specific research contents are as follows:First,the complex geological structure leads to complex wave group features on deep-reflected seismic sections.The single and fixed basis functions or the simple linear combination of various basis functions cannot most effectively characterize the internal structural features of the data.Based on the theory of MCA and compressed sensing,a weighted MCA sparse representation method is proposed according to the sparseness of deep seismic data in each sparse domain.The prior information is used to constrain the weights of various sparse transform dictionaries in their sparse representation,and the sparsest description of complex seismic data is realized.The regularized reconstruction of deep seismic data is produced by combining corresponding threshold parameters and iterative algorithms.Then,the reconstruction quality of seismic data with irregular complex structures is improved.Secondly,aiming at the problem that the difference between deep effective weak signal and noise interference band is small and difficult to distinguish,a weak signal extraction method for OVT domain is proposed based on the theory of compressed sensing.It uses the information related to noise intensity to constrain the inversion process,overcomes its dependence on signal sparsity,and achieves effective extraction of weak signals.In order to further improve the weak signal extraction effect of complex geological structures,the CEEMD method is introduced to modally decompose the signal according to the characteristic information of the signal itself.Then the components with more noise are determined by cross-correlation analysis,and CS denoising is performed in these components.The enhancement operator is introduced in the denoising process to ensure that the effective information is enhanced while the noise is not,and the weak signal extraction effect is further improved.Finally,the low-frequency information can reflect the basic trend of the formation,improve the accuracy of velocity analysis and the imaging accuracy of deep structures.The low-frequency compensation method of seismic data based on compressed sensing theory can compensate the low-frequency components reasonably,thus improving the data quality.However,this method is greatly affected by noise.This paper comprehensively uses the logging data to constrain the reflection coefficient inversion process.It can improve the anti-noise ability in the inversion process,improve the inversion precision of the reflection coefficient,and improve the compensation effect of low-frequency information,thereby enhancing the weak effective signal in seismic data.Compressed sensing theory plays an important role in improving the quality of seismic data.In this paper,the correctness and effectiveness of the proposed method are verified by numerical experiments and actual data tests,which provides a new idea for improving the regularity,signal-to-noise ratio and fidelity of seismic data.