Several New Digital Techniques And Its Application In Seismic Prospecting

The high signal-to-noise rate (SNR) and resolution is the aim that people arepursuing for gaining good seismic profile and geologic interpretation. Thegeneral media filter(MF) , weighed media filter(WM) , multistage mediafilter(MLM) and self-adaptive weighed media filter(AWM) can eliminaterandom noise effectively in the seismic data. The least mean square (LMS) filterin frequency domain can remove refraction effectively in the seismic data. Thesurface wave can be wiped of by e 2-D wavelet transform. Adjusting the weightcoefficient and the length of filter can compress wavelet and heighten the seismicdata revolution. Crossing-zero wavelet can use the crossing-zero point to checkthe point of variety acutely in the seismic signal, and identify the folium in theseismic data to improve the revolution.Adaptive LMS filter in frequency domain is that the signal can be changedto the field of frequency firstly, and adopt some transforms to adjust the filtercoefficient. There are three merits: firstly, the count of process the data decreasesgreatly comparing with count in time domain. Secondly, the processing ofself-adaptive convergence can be improved. Thirdly, the processing in frequencydomain can easily divide into different parts. This paper apply the adaptive LMSfilter to process the seismic data, according the model and practice data, we cansee that this method has good effect and value of application.The median filter is simplex and effective method, it's extensively applied indigital processing domain etc. Although it is only thirty years since the 1970'swhen L.R.Rabmer, M.R.Sambur and C.E.Schmidt brought forward the originaltheory of median filter, the theory and applied method of the median filterdevelop fast during this period. The median filter is well known for its noisesuppression and ability to preserve edges. On the early time, the study of themedian filter is restricted to signal and picture processing of electric technology.In recent years, the method has been applied in the seismic exploration, forexample the median filter is able to attenuate cohere wave field, attenuate thedowngoing wave in VSP and attenuate surface wave in reflected seismicinformation. Fitche and Hardage applied the median filter in velocity filtering ofmultitrace VSP, which occurred good effect. Xu Changlian applied MF in theprocessing of earth surface seismic data, which supplied a new mean forimproving SNR of pre-stack data. Along with the MF theory development, somenew means like the weighted median filter, the multistage median filter and theadaptive multistage median filter etc. reinforce the MF technology. Some newalgorithm like quick median filter algorithm, median filter with transversalrecursion and abstract class of median filter algorithm make median filter toapplied to practice better.The median filter is a kind of special non-linear filter, it is different fromlinear filter. The amplitude spectrum and phase spectrum of the linear filtercompletely determine the characters of linear filter in frequency field and timefield. On the compare of it, the response of the median filter to pulse is zero,there are no 'real' amplitude spectrums and phase spectrums, so although thetheory of the median filter is very simple, but the specialities of the median filterare more difficultly to be understood than linear filter. This study analyzes andsummarizes the characters of median filter through the method of the twodifferent median filters responding to existing signals. We summarize somecharacters as:â‘  The median filter can distort the signal shape, and the degree of distortionis relevant to the length of filter;â‘¡ The median filter can lead to false-high-frequency, so the signal after MFshould run low-frequency-pass filtering according to conditions;â‘¢ The median filter can attenuate unsymmetrical signal with phasedistortion, the situation with the weighted median filter is more serious, so wemust affirm that the phase distortion is allowed or not before the median filter isused;â‘£ Although1-D WM can improve the resolution after choosing properweight coefficient, it can bring 'little step' effect. So we recommend make asmooth processing after WM.Although the median filter technology is very important on the signalprocessing, especial on the unstable signal processing, one of the severedeficiencies is that it can arouse 'minute' destroy and lose of signal detailstructure. On the picture processing, the imperfect of MF is more serious than1-D signal processing. The reasons are come from two aspects: firstly, 2-D signalhas no base signal, that is mostly all 2-D signals would be destroyed withdifferent degree after MF;secondly, the detail structures like filaments andcorners always contain important information, the destroy and lose is moreunacceptable than noise. So the study of median filters with protecting detailsbecomes one of the important aspects of non-linear filter studies. The multistagemedian filter has noise suppression and ability to preserve edges at the same time,so it is what we are looking for.This study applies 1-D median filter and 2-D multistage median filter to theactual 2-D seismic profile data, and we analyze the conclusion, consider that 2-Dmultistage median filter have the ability of protecting detail, but its ability ofnoise suppression is low.On high accuracy seism-geology position calibration, 1-D MF can improveinterlayer information through decreasing the energy of dominant strata, but itgives rise to 'platform' effect;WM can avoid 'platform' effect and improveresolution, but because its attenuation ability is too strong, some usefulinformation disappear. So when we apply MF to composite recording, we shouldchoose the style and filter length. When improving interlayer information, weusually choose MF, and when we will improve resolution, WM is more suitable.The filter length should be determined according to the filter purpose too. Thestudy shows that the composite recording MF with 7 points and WM with 5points is much better.After this study we find that MF is different from others filters, it can't bedescribed as certain parameters, the experiment is the best way to choose theparameters of MF according to practical situation, only if this, the method cangives play to action.The general MF needs many experiments for adjusting the window lengthof filter, because the window length of filter is changeless in processing of filter.We bring forward the adaptive-weighed median filter that can adjust thewindows length of filter with real number weight coefficient can eliminaterandom noise in the precondition of protect useful signal. Adaptive processingcan protect useful information effectively, and weight coefficient can eliminatethe random noise availably. In the processing, adaptive-weighted median filteronly need to choose the fiducial median filter coefficient to process adaptivelythe all course. Through processing the practice seismic data and analyzing itsfrequency spectrum, we know that it can eliminate random noise effectively.In 1986, the research of wavelet comes to climax, but there are many yearsusing idea of translation and flex in the wavelet theory. Morlet ( France) advancethe wavelet theory in analyzing the seismic data, and advance the geometricalsystem of wavelet with Grassmann(France). Wavelet is often compared tomathematics microscope, and can magnify, reduce and translate the signal. It canstudy the signal characters through checking the changes of differentamplificatory rate.Today, wavelet transform is applied to seismic data collection,processing,inversion and interpretation. This paper uses wavelet transform to enhance theSNR and revolution of the seismic data in the seismic prospecting. The analysisof the time-frequency based on wavelet has no the limit of time-frequency, canoffer very great frequency spectrum decomposability. The 2-D wavelet transformcan convert the seismic data into them in frequency domain, and have good partcharacters in the frequency. The refraction and surface wave can be distinguishedfrom the seismic data. Thus, the useful wave and other waves can be detachedwith their differences of time-frequency. From this paper, 2-D wavelet transformcan eliminate greatly surface wave in the seismic prospecting.Otherwise, the wavelet transform can be used to enhance the revolution ofseismic data. In the seismic prospecting, searching the singularity of qualitativeparameter of crustal substance is very significative, judging the size and positionof the singularity can interpretation the abnormity. The wavelet transform has thelocal of time and frequency, so it is an effective method to describe and check thesingularity of function. Using the singularity (e.g., crossing-zero point) ofwavelet transform of signal to express the signal, we can to enhance theresolution of seismic. This study checks the singularity of signal with thecrossing-zero point of wavelet, and identifies the folium of seismic data toimprove the resolution.From this study, we can know that there are many methods of high SNR andresolution, the methods in this paper can obtain good results. At the same time,we should study the choosing of parameter more to obtain better methods thatcan get good results.