Super Wavelet Sparse Representation And Study Of Coherent Technology

The development and utilization of natural energy greatly promote the development ofsociety, increase the social productive forces, bring great convenience for people’s life. Itmust be as large degree of rational development and utilization of these natural treasuresunderground as possible in order to realize sustainable use of natural energy particularly theuse of underground energy such as coal, oil, natural gas. With the growing demand for energy,underground energy exploration becomes more important, while finding new oil and gasreserves is becoming increasingly difficult, then it requires people to look for oil and gasstorage areas more fast, more accurate and more efficient. Seismic exploration is an importantmethod of analyzing underground geological structure, seismic data includes a variety ofseismic attributes such as amplitude, waveform, frequency, attenuation, energy whichimplicating a wealth of information, the interpretation of seismic attributes can effectivelysolve the geological structure problems, especially in favor of searching for oil and gasreservoirs. Fault interpretation is one of the most important seismic interpretation problems.Accurate identification of fault lines can help researchers make correct analysis and judgmentof underground structures and oil and gas reserves, so it has important practical significancein improving the accuracy of fault lines identification.Wavelet can capture singular point of one-dimensional signal powerfully, but it is not sosatisfactory in capturing high-dimensional signal bizarrerie, while the super-wavelet can solvethis problem. In recent years, the use of super-wavelet has been used more and more widely inthe field of seismic exploration. Surfacelet transformation refers to decomposing the signalusing multi-resolution first, and then use the multi-dimensional directional filter banks totransform coefficients in the same direction on the merger, Surfacelet transform caneffectively capture characters of high-dimensional signal. It is convenient for discretethree-dimensional signal processing, so it is very suitable for processing three-dimensionalseismic data to improve signal to noise ratio. In this paper we use surfacelet transform toprocess seismic data, taking full advantage of anisotropic character. We have achieved goodresults on the three-dimensional seismic data processing, which provides more help forfollow-up to extract fault information.The basic principle of the coherent method refers to calculating the degree of similaritybetween each sampling point in every seismic trace and adjacent seismic trace in a certaintime window in the whole three-dimensional seismic data, the degree of similarity can berepresented as a specific value which finally form a new three-dimensional data volume, and data in the volume represents degree of correlation between the point and other points inneighborhood. When the degree of correlation is large, the value is large, when the coherencevalue is large, the value is small, this result reflects the discontinuity among seismic traces,which contribute to identify faults, fractures, special rock and so on.The main contents of this paper include the following points:1. Research of seismic exploration knowledge, including seismic data acquisition,seismic data processing and surfacelet transformation of sparse representation theory of theimage. We have a comprehensive and in-depth study, including origin of the direction filter(DFB), specific concepts and operational processes of multidimensional directional filter(NDFB), and a detailed study and analysis of surfacelet;2. Seismic data denoising method based on Surfacelet transform. We apply surfacelettransform in target area of three-dimensional seismic data and then we get seismic datasub-bands in different directions under different scales. These sub-bands are formed by aseries of factors. We process the coefficients of the fine-scale layer since the noise of seismicdata are often distributed in the high frequency part, and data will be reconstructed aftercoefficient processing to improve the signal to noise ratio of seismic data purposes;3. Study and research of the main techniques used to fault identification in seismicexploration, including the first-generation coherent technology based on cross-correlation, thesecond generation coherent technology based on similarity, the third generation coherenttechnology based on eigen structure, in addition we have analysis and summary of coherencetechnology application in fault identification;4. Improved coherent algorithm based on multi-eigenvalues. On the basis of thethird-generation coherent technology, improve the method of calculating coherent values ofeach point in seismic traces by using multiple eigenvalues to improve the degree of energymaintained in seismic data. Experimental results show that compared with the traditionalmethod(C3), improved multi-eigenvalues coherence algorithm is more suitable for theextraction of fault information, in addition, when combined with Surfacelet transformation,the effect is more pronounced.