Research And Application Of The Spectrum Imaging On Seismic Data

Use spectrum imaging technique analyzes the seismic information has practical significance because the seismic information about gas and oil is the first-hand information obtained in the process of reservoir exploration. Spectrum imaging technique, the core of which is time-frequency analysis, is a reservoir prediction and interpretation tool in the frequency domain. Firstly we discuss the question of the seismic information optimization, then summarize several common time-frequency analysis methods, and research the application of them on the thin layer, sedimentary cycle, gas sand detection and on the combination of them with the methods of information optimization.At the aspect of seismic information optimization, we research the effects of common used methods of wavelet threshold shrink and generalized S transform time-frequency filtering. Then introduce the multi-scale geometric analysis(MGA) methods which widely used in the image processing domain recently to the seismic random noise suppression domain, and summarize their basic theories, development, advantages and disadvantages, mainly research the Curvelet transform and Contourlet transform, discuss the influence of threshold functions to the result. On the basis of this, we apply the information optimization methods above to random noise suppression of synthetic seismogram model with different signal noise ratio, and find that MGA methods have better effects compare with the wavelet transform both in the qualitative and quantitative aspects.Starting from the limitation of Fourier transform, we summarize several common time-frequency analysis methods, which include short time Fourier transform, wavelet transform, wavelet packet transform, S transform, generalized S transform and Wigner-Ville distribution, and analyze the merits and demerits of them. Mainly on the generalized S transform, discuss its time-frequency resolution on the theory via the research of its window parameters. And use the LFM signal to detect the time-freqency resolution of several time-frenquency analysis methods.On the basis of the theory, we discuss the application of spectrum imaging technique through synthetic seismic records and practical seismic records. (1) discussing the characteristic of the reflection coefficient sequence and time-frequency spectrum of thin layer, on the basis of which, we achieves the thin layer identification through the single frequency profile analysis. (2) establishing models of positive, negative and combined cycles and discussing their time-frequency spectrum characteristics, on the basis of which to divide the seismic cycles. (3) discussing the impact of random noise on time-frequency analysis and combining generalized S transform and the methods of seismic information optimization to analyze the pre-stack seismic data for obtaining more accuracy information. (4) discussing the spectral decomposition attributes of seismic data and the application in the identification of layer and breakpoint. (5) discussing the application of the generalized S transform to the gas sand low shadow detection through practical 3D data analysis.