Time - Frequency Analysis Method And Its Application In Seismic Data Processing

The author using the linear time-frequency theorem such as the short-time Fourier transform, wavelet transform and S transform (include generalized S transform) computes time-frequency parameter. According to the computed time-frequency parameter result, the author not only develops the method to increase the resolution and signal-noise ratio, but also extract the instantaneous parameter and determine the depth variety of thin layer.In this paper, the author firstly introduce the basic idea and the mainly mathematics tool on the non-stationary signal and time-frequency analysis method. The author also analysis the relationship of each linear time-frequency analysis method and compare their advantage and disadvantage.Using the time-frequency method of wavelet transform, the band-width of frequency broadens, so the seismic signal resolution can be increased. The author make full use of the advantage of wavelet transform method and spectrum equibalance method, combine the two method to process the result of different scale using spectrum equibalance, the resolution can be increased. According to the result of theory data and real data, we can see the method is very useful in increasing the band-width of frequency and the resolution has been increased. Through using the method in real data, it can be proven the wavelet transform of spectrum equibalance method is very useful and reliable.In the basis of analyzing the character of different linear time-frequency method, a new method of time-frequency filter is developed to improve the signal noise ratio based on the generalized S transform method. Firstly, using generalized S transform method to compute the time-frequency profile of each trace and then combining the result of 1-D spectrum analysis and F-K analysis, at last determining the time-space filter factor on the time-frequency profile. After these step, using generalized S inverse transform method to get the filtered profile. In this paper, the author shows some example to verify this method is very effective, such as in suppressing the noise of tube wave and surface wave. Comparing with 1-D filter, the generalized S transform filter method can filter the noise that has the similar frequency with the effective signal, such as the tube wave. Comparing with F-K filter, the generalized S transform filter method can filter the noise that has the similar apparent velocity with the effective signal. In addition, the generalized S transform method (include generalized S transform) not only can be used with time-domain algorithm but also can be used with frequency-domain algorithm. For saving the running time,...