The Application Of The Modified Wiener Filtering In The Seismic Data Processing

The seismic prospecting is one method of the geophysics exploration. Every geophysics method is based upon the research of a certain physics characteristic of rocks, and the basis of the seismic prospecting is the elasticity of the rocks. The seismic prospecting uses the artificial source to produce elastic wave, then measure the vibration of the ground with the seismic detector in the different lines. Usually record the data in the memorizer with the digital type to improve the signal-to-noise(S/N) ratio and obtain the useful information using the processing of the computer. Show the results in the type with which easily carrying on the geology explanations. Because the path, the vibration strength and wave form of the seismic wave vary with the different elastic characteristics and the different geometrical paths in its propagation course in the medium, if we master these variety regulations, according to the traveling time and the information of velocity, can predict the propagation path and the structure of the medium; And based on the parameters, such as the amplitude, frequency and geologic layer's speed etc., can infer the characteristics of the rocks to carry out the aim of the prospecting.With the continuous research of the seismic prospecting, aiming at the target of the exploration pay attention to the deeper layer's structures. However, because the goal layer is deeper, the singal of the seismic wave produced by the surface source decays as a result of the long distance propagation and diffuseness and the nonelastic effect, and the reflected signal received on the surface is weaker. Under the background of the random noise, the seismic reflected signal sometimes disappears and appears, even thoroughly disapears in the noise, which make the same axis of the reflection wave hard to trace and the exploration is harder. How to improve the resolution ability and S/N ratio becomes the key problem of the seismic data processing. For this question, people developped various filter techniques and successfully applied in the seismic data processing to improve the S/N ratio.In the seismic prospecting, not only record various information of the reflection wave but also record the information of the various noise waves. The differences of the useful signals and the noise wave represent many spects (e.g. spectrum, the propagation direction, energy etc.).At present, in the seismic data processing, the filter is carried out by either the modulation electricity filter or the digital filter. The method using the spectrum characteristics to suppress the noise waves and give prominence to the useful signals is the digital filter. The digital filter, using the calculations of the digital electrical computer, carrys out the filter.While carrying on the digital filter, first digitize the continuous signal, so the data inputted and outputted are the digital data.In the seismic exploration, it is much more important to suppress the noise wave and improve the S/N ratio. The technique in the data processing is the same as the techniques in the data collection, which is the differences between the useful signals and the noise. The digital filter using the differences of the frequency and the apparent velocity between them, suppresses the noise, and the methods are separately called frequency filter and apparent velocity filter. A basic problem of the filter research is: How to design and build up the best or optimal filter. The optimal filter is the one designed according to a certain best standard. Suppose the input of the linear filter are the summation of the useful signals and the noise, and they both are general steady processes and their statistical characteristics have been known. According to the least average square error standard, which is that the square value of the difference between the output signal and the expecting signal is the least, obtain the parameters of the optimal linear filter. This kind of filter is called the Wiener filter.As being the theoretic least average square error linear filter, the application of the Wiener filter needs to solve the Wiener—Hopf equation. In the solution the mass calculations make the application of the Wiener filter restricted. In fact usually use the extendible type .Because the power spectrum density of the signal P_ss(w) and noise P_vv(w) are unknown beforehand, a method of solution is to obtain the estimative value of the power spectrum density of the signal (P|^)_ss(w) , or obtain the power spectrum density of the noise P vv( w) with a certain method.Under the condition of the noise estimate obtained, the improved Wiener filter can obtain the filter parameters according to the power spectrum of the signal observed .for examining the effect of the noise wave suppressed of the Wiener filter, firstly design the sine signal with the noise wave, and carry on the experiment and compare the result with the one in the Wiener2 filter in the Matlab to obtain the better filter effect. Secondly the short pulse vibration which is produced by the source, propagated and received by us on the ground or in the well is called vibration wavelet. It can be comprehended that it is a signal which has certain starting time and limited energy and can decay in a short time. for the application in the seismic data processing, select the Ricker wavelet usually used in the seismic wave, with adding the Gaussian noise, then obtain the model signal record with the noise. In the experiment, under the condition of certain filter parameters, enlarge the power of the noise to examine the ability of the noise suppressed of the Wiener filter. At last, in the many channel seismic data processing, make use of the same source point records to carry on the experiment. Because of the random noise, hardly observe the obvious same phasic axis, according to the estimation of the random noise, can obtain the improved Wiener filter. Using the improved Wiener filter, the effect makes the same phasic axis more obvious and noise weakened. Following, when enlarging the noise and retaining other conditions, the effect show that the result isn't like the one expected when the noise is smaller. However under the same condition, the effect compared with the Wiener2 filter in the Matlab has a great improvement.With the many experiments, the improved Wiener filter has the brief, convenience and easier computation characteristics. While carry on the seismic data processing, under the condition of the estimation of the random noise known, obtain the best filter parameters with adjusting the increasing factors to improve the S/N ratio of the signal. In the actual information, for complicated noise, it can improve the application of the improved Wiener filter when the estimation of the noise is confirmed. In the experiment the datum is the single section of the same phasic axis, and in the actual signal exist many sections of the same phasic, so need to do a research to the condition of many same phasic axis and examine the actual feasibility.