Study On Sparse Representation And Frequency Expanding Processing Method Of Seismic Data

With the development of exploration and exploitation,the conventional seismic data processing methods are no longer suitable for the exploration of complicated reservoirs with high difficulty and high precision,such as stratum and lithology reservoirs.Especially now the exploration of oil and gas reservoir is gradually dominated by thin layer exploration.The vertical resolution of conventional seismic data processing and attribute extraction technology is not enough to distinguish the thickness of these layers.It is customary to call the reservoir with thickness less than the tunable thickness and the top and bottom interfaces on the seismic section unidentifiable as thin layer.The seismic resolution is closely related to the frequency bandwidth of reflected signals.Therefore,broadening the seismic frequency band is of great significance to improve the seismic resolution,which is also one of the key problems to be solved in high-resolution seismic exploration.Based on the theory of sparse representation and compressed sensing,this paper uses two sparse decomposition algorithms,matching pursuit and basis pursuit,to carry out sparse decomposition of seismic data and obtain the sparse representation of seismic data,so as to broaden the seismic frequency band and improve the resolution of seismic data.The matching pursuit expands the seismic signal into a series of superposition of the best matching wavelet,and extracts the thin layer thickness information by combining the thin layer seismic response rule,which is another point of view to break through the limit of the traditional seismic resolution.Two sparse decomposition algorithms,matching pursuit and basis pursuit,are used for sparse representation of seismic data.When building the sparse dictionary of matching pursuit,the over-complete wavelet library built by the traditional way is huge,and the amount of search and calculation based on it is huge.As a result,the calculation efficiency of the whole process is low,which cannot meet the needs of practical application.The computational efficiency of the optimized Complex Field Fast Matching Pursuit algorithm is greatly improved,which fully meets the needs of practical applications.The building of the target function of seismic data inversion based on the basis pursuit algorithm is also the building of sparse dictionary.Using odd-even decomposition theory construct the objective function in time domain,the sparse dictionary is very cumbersome,which can lead to inversion problem such as low inefficient operation and unstable results.So we try to construct the objective function in the frequency domain,and do the inversion based on basis pursuit algorithm.This method is called SineCosine Synergistic Inversion Based on Basis Pursuit,which is efficient and stable.Theoretical analysis and practical seismic data trial show that the sparse representation of seismic data obtained by matching pursuit and basis pursuit algorithm can effectively broaden the frequency bandwidth of seismic data and improve the seismic resolution.The anastomosis rate between the predicted thin layer thickness information and actual well data is high.