F-K域地震道插值方法研究
|
摘要
随着油气勘探开发的不断深入,对地下构造研究的精度要求越来越高。在地震勘探中,由于工作量的限制及施工条件的影响等,常常使某些测线方向上地震记录的道间距较大。这就造成了空间采样率的严重不足,对F-K滤波、Radon变换滤波,尤其是偏移成像等处理会产生严重的影响。偏移剖面上出现空间假频及频散现象,会使横向分辨率变差,使解释人员无法正确解释地下构造。解决空间采样率不足的最好办法是在野外采集时直接减少空间采集间隔,但这种方法受到客观条件的限制。当不能在野外采集直接减少空间采集间隔时,采用地震道插值技术来减小空间采集间隔就成为比较好的方法,该方法不能增加信息量,可以克服空间假频。
地震道插值可以用来加密空间采样率,防止偏移时频散的出现,提高信噪比。对一些要求作精细地质解释的地区,只需对原先的地震资料作地震道插值,然后进行偏移处理,而不必重新进行施工采样。这样不但节约了大量人力、物力,而且还可大大缩短生产周期。经过地震道插值后的数据体增加一倍,诸如速度、振幅、频率等地球物理信息更加丰富。地震道插值基本上可以起到野外密集采样所达到的地质效果。由于地震道插值克服了空间假频,使其地球物理信息更加真实地反映地下地质体的地球物理特征,更有利于进行构造解释和地震地层学的研究。
因此国内外许多学者都关注这个研究课题。国外一些公司,如西方地球物理公司(WGC)、HGS、CGG等已把地震道插值引人常规处理流程。地震道插值已经成为地震资料常规处理中不可缺少的步骤。早期的地震道插值方法是在滑动时空窗内扫描同相轴的倾角,然后沿若干倾角方向加权来产生内插地震道。这些方法都需要对同相轴进行分离,所以需要反射倾角的先验信息。1989年Spitz提出了不用分离同相轴的地震道插值方法,随后又有很多科学家提出不用预先知道同相轴倾角的地震道插值方法。如Claerbout提出的T-X预测误差滤波插值方法和Gulunay提出的抗假频F-K域地震道插值方法(UFKI)等,这些方法也成为现在常用的地震道插秩??但这些方法都存在它们各自的问题。其中Sinc地震道插值方法虽然在进行地震道插值时速度快、易于实现,但是无法正确内插具有空间假频的地震道。Spitz的F-X域地震道插值方法和Claerbout的T-X域预测误差滤波插值方法等虽然可以正确内插具有空间假频的地震道,但是这些方法计算量特别大,使其在实际数据中的可用性大大降低。Gulunay提出的抗假频F-K域地震道插值方法(UFKI)虽然可以正确内插具有空间假频的地震道,且运算速度快、效率高。但是其插值系数不能为任意整数,也不能对三维地震数据进行地震道插值,这就使它的实用性大大降低。
根据现有这些方法的问题,本文在介绍了常用Sinc地震道插值方法,Clearout的T-X域预测误差滤波插值方法,Spitz的F-X域地震道插值方法及抗假频F-K域地震道插值方法(UFKI)的基础上。着重分析研究了另一种F-K域地震道插值方法——广义F-K域地震道插值方法(GFKI)。该方法可以正确内插具有空间假频的地震道,内插出的地震道波形自然,且具有计算量小、效率高等优点。广义F-K域地震道插值方法(GFKI)的插值系数L可以为任意整数,即两相邻地震道之间可以被插入任意整数的插值道。该方法也可以被应用于三维地震数据中。本文通过实现广义F-K域地震道插值方法(GFKI),分析了该方法在进行实际地震道插值时会出现的一些具体问题,并提出了解决方案。讨论了F-K域地震道插值的GFKI方法和UFKI方法的区别。最后通过理论模型试算与实际资料处理验证了,广义F-K域地震道插值方法(GFKI)比其他常用地震道插值方法具有更高的效率。
Along with oil-gas exploration development unceasingly, the research on the accuracy of the underground structures have become increasingly demanding. In the seismic survey, as a result of the work load limit and execution conditions and so on, the record samples interval is frequently big in survey. This has lead to the serious insufficiency of spatial sampling .The F-K filter, the Radon transformation filter, in particular such as offset processing will have a serious impact. The offset section will have phenomenon of spatial aliasing and frequency dispersion, the horizontal resolution will deteriorate. So that will cause to explain the subsurface structure incorrectly. The best way to resolve the spatial sampling problem is directly reduce spatial samples interval in the field, but this approach by objective conditions.when this approach can not be applied , seismic trace interpolation techniques is a good method to resolve the spatial sampling problem. This method can not increase the information, spatial aliasing can be overcome.
Seismic trace interpolation can be used to encrypt the spatial sampling, prevent the emergence of frequency dispersion, enhance signal noise ratio. Sometimes we only do seismic trace interpolation to the original seismic data, and then do the offset process, without having to re-construction sampling to the geological regions which required high accuracy. This will not only save a great deal of manpower and material resources, but also greatly shorten the production cycle. The data after seismic trace interpolation are twice as before. ?? as speed, amplitude, frequency and other geophysical information richer. Seismic trace interpolation basically may have the geological effect which intense the sampling in the field can achieve. As seismic trace interpolation overcomes the spatial aliasing, geophysical information could more truly reflect the underground geological characteristics. So it is more advantageous in carrying on the structure explanation and the seismic stratigraphy research.
So many scholars at home and abroad are more concerned about this study. Some foreign companies, such as the Western Geophysical Company (WGC), HGS, CGG using seismic trace interpolation conventionally. Seismic trace interpolation of seismic data has become indispensable to the conventional process. The early seismic trace interpolation scan the events dip in space-time window,and then along the dip direction of a number of weighted to do seismic trace interpolation. These methods are necessary to separation the events, so they need a priori information of events dip. In 1989 Spitz proposed a method that dose not need to separate events, afterward a lot of scientists to propose seismic trace interpolation that did not need to know the events dip. Claerbout proposed T-X domain prediction error filtering interpolation and Gulunay proposed unaliased F-K domain trace interpolation (UFKI). These methods are now used commonly, but these methods also have their problems. Though the speed of Sinc interpolation method do the interpolation is quick and easy to realize, this method can not to interpolate the date with spatial aliasing correctly. F-X trace interpolation method of Spitz and T-X domain prediction error filtering interpolation method of Claerbout could interpolate the data with spatial aliasing correctly, but these methods need large amount of calculation, so it could reduce the usability in the field data greatly. Gulunay proposed unaliased F-K domain trace interpolation (UFKI).This method ??d interpolate the data with spatial aliasing correctly, and the operating speed is quick, the efficiency is high. But its interpolation coefficient cannot be the random integer, and it also cannot carry on the 3D data record. This greatly reduced its usability.
According to the questions of these methods, this article introduces the commonly used seismic trace interpolation methods such as Sinc interpolation, T-X domain prediction error filtering interpolation proposed by Clearout, F-X seismic trace interpolation proposed by Spitz and unaliased F-K domain trace interpolation (UFKI). Then emphatically analytical another F-K seismic trace interpolation method, that is the generalized F-K seismic trace interpolation (GFKI).This method could interpolate the data with spatial aliasing correctly ,the interpolated trace are profile nature, the operating speed is quick and efficiency is high.The interpolation coefficient L of generalized F-K seismic trace interpolation (GFKI) could be the random integer, that means between two neighboring trace insert the number of the interpolation trace could be random integer. This method may also apply in the 3D data record. Through realizes the generalized F-K seismic trace interpolation (GFKI), this article analysis the specific issues when carrying out the method in field data and proposed solutions for this issues. Then I discussed the difference between GFKI and UFKI. Finally many kinds of theoretic data models designed and real seismic data are used to verify the conclusion that generalized seismic F-K interpolation domain (GFKI) are more efficient than other common seismic trace interpolation methods.
地震道插值可以用来加密空间采样率,防止偏移时频散的出现,提高信噪比。对一些要求作精细地质解释的地区,只需对原先的地震资料作地震道插值,然后进行偏移处理,而不必重新进行施工采样。这样不但节约了大量人力、物力,而且还可大大缩短生产周期。经过地震道插值后的数据体增加一倍,诸如速度、振幅、频率等地球物理信息更加丰富。地震道插值基本上可以起到野外密集采样所达到的地质效果。由于地震道插值克服了空间假频,使其地球物理信息更加真实地反映地下地质体的地球物理特征,更有利于进行构造解释和地震地层学的研究。
因此国内外许多学者都关注这个研究课题。国外一些公司,如西方地球物理公司(WGC)、HGS、CGG等已把地震道插值引人常规处理流程。地震道插值已经成为地震资料常规处理中不可缺少的步骤。早期的地震道插值方法是在滑动时空窗内扫描同相轴的倾角,然后沿若干倾角方向加权来产生内插地震道。这些方法都需要对同相轴进行分离,所以需要反射倾角的先验信息。1989年Spitz提出了不用分离同相轴的地震道插值方法,随后又有很多科学家提出不用预先知道同相轴倾角的地震道插值方法。如Claerbout提出的T-X预测误差滤波插值方法和Gulunay提出的抗假频F-K域地震道插值方法(UFKI)等,这些方法也成为现在常用的地震道插秩??但这些方法都存在它们各自的问题。其中Sinc地震道插值方法虽然在进行地震道插值时速度快、易于实现,但是无法正确内插具有空间假频的地震道。Spitz的F-X域地震道插值方法和Claerbout的T-X域预测误差滤波插值方法等虽然可以正确内插具有空间假频的地震道,但是这些方法计算量特别大,使其在实际数据中的可用性大大降低。Gulunay提出的抗假频F-K域地震道插值方法(UFKI)虽然可以正确内插具有空间假频的地震道,且运算速度快、效率高。但是其插值系数不能为任意整数,也不能对三维地震数据进行地震道插值,这就使它的实用性大大降低。
根据现有这些方法的问题,本文在介绍了常用Sinc地震道插值方法,Clearout的T-X域预测误差滤波插值方法,Spitz的F-X域地震道插值方法及抗假频F-K域地震道插值方法(UFKI)的基础上。着重分析研究了另一种F-K域地震道插值方法——广义F-K域地震道插值方法(GFKI)。该方法可以正确内插具有空间假频的地震道,内插出的地震道波形自然,且具有计算量小、效率高等优点。广义F-K域地震道插值方法(GFKI)的插值系数L可以为任意整数,即两相邻地震道之间可以被插入任意整数的插值道。该方法也可以被应用于三维地震数据中。本文通过实现广义F-K域地震道插值方法(GFKI),分析了该方法在进行实际地震道插值时会出现的一些具体问题,并提出了解决方案。讨论了F-K域地震道插值的GFKI方法和UFKI方法的区别。最后通过理论模型试算与实际资料处理验证了,广义F-K域地震道插值方法(GFKI)比其他常用地震道插值方法具有更高的效率。
Along with oil-gas exploration development unceasingly, the research on the accuracy of the underground structures have become increasingly demanding. In the seismic survey, as a result of the work load limit and execution conditions and so on, the record samples interval is frequently big in survey. This has lead to the serious insufficiency of spatial sampling .The F-K filter, the Radon transformation filter, in particular such as offset processing will have a serious impact. The offset section will have phenomenon of spatial aliasing and frequency dispersion, the horizontal resolution will deteriorate. So that will cause to explain the subsurface structure incorrectly. The best way to resolve the spatial sampling problem is directly reduce spatial samples interval in the field, but this approach by objective conditions.when this approach can not be applied , seismic trace interpolation techniques is a good method to resolve the spatial sampling problem. This method can not increase the information, spatial aliasing can be overcome.
Seismic trace interpolation can be used to encrypt the spatial sampling, prevent the emergence of frequency dispersion, enhance signal noise ratio. Sometimes we only do seismic trace interpolation to the original seismic data, and then do the offset process, without having to re-construction sampling to the geological regions which required high accuracy. This will not only save a great deal of manpower and material resources, but also greatly shorten the production cycle. The data after seismic trace interpolation are twice as before. ?? as speed, amplitude, frequency and other geophysical information richer. Seismic trace interpolation basically may have the geological effect which intense the sampling in the field can achieve. As seismic trace interpolation overcomes the spatial aliasing, geophysical information could more truly reflect the underground geological characteristics. So it is more advantageous in carrying on the structure explanation and the seismic stratigraphy research.
So many scholars at home and abroad are more concerned about this study. Some foreign companies, such as the Western Geophysical Company (WGC), HGS, CGG using seismic trace interpolation conventionally. Seismic trace interpolation of seismic data has become indispensable to the conventional process. The early seismic trace interpolation scan the events dip in space-time window,and then along the dip direction of a number of weighted to do seismic trace interpolation. These methods are necessary to separation the events, so they need a priori information of events dip. In 1989 Spitz proposed a method that dose not need to separate events, afterward a lot of scientists to propose seismic trace interpolation that did not need to know the events dip. Claerbout proposed T-X domain prediction error filtering interpolation and Gulunay proposed unaliased F-K domain trace interpolation (UFKI). These methods are now used commonly, but these methods also have their problems. Though the speed of Sinc interpolation method do the interpolation is quick and easy to realize, this method can not to interpolate the date with spatial aliasing correctly. F-X trace interpolation method of Spitz and T-X domain prediction error filtering interpolation method of Claerbout could interpolate the data with spatial aliasing correctly, but these methods need large amount of calculation, so it could reduce the usability in the field data greatly. Gulunay proposed unaliased F-K domain trace interpolation (UFKI).This method ??d interpolate the data with spatial aliasing correctly, and the operating speed is quick, the efficiency is high. But its interpolation coefficient cannot be the random integer, and it also cannot carry on the 3D data record. This greatly reduced its usability.
According to the questions of these methods, this article introduces the commonly used seismic trace interpolation methods such as Sinc interpolation, T-X domain prediction error filtering interpolation proposed by Clearout, F-X seismic trace interpolation proposed by Spitz and unaliased F-K domain trace interpolation (UFKI). Then emphatically analytical another F-K seismic trace interpolation method, that is the generalized F-K seismic trace interpolation (GFKI).This method could interpolate the data with spatial aliasing correctly ,the interpolated trace are profile nature, the operating speed is quick and efficiency is high.The interpolation coefficient L of generalized F-K seismic trace interpolation (GFKI) could be the random integer, that means between two neighboring trace insert the number of the interpolation trace could be random integer. This method may also apply in the 3D data record. Through realizes the generalized F-K seismic trace interpolation (GFKI), this article analysis the specific issues when carrying out the method in field data and proposed solutions for this issues. Then I discussed the difference between GFKI and UFKI. Finally many kinds of theoretic data models designed and real seismic data are used to verify the conclusion that generalized seismic F-K interpolation domain (GFKI) are more efficient than other common seismic trace interpolation methods.
引文
[1]陆基孟.地震勘探原理(上,下册)[M].北京:石油大学出版社,1993.
[2]黄德济,贺振华,包吉山.地震勘探资料数字处理[M].北京:地质出版社,1990.
[3]何樵登.地震勘探原理和方法[M].北京:石油大学出版社,1985.
[4]牟永光.地震勘探资料数字处理[M].北京:石油工业出版社,1981.
[5] Claerbout J. EARTH SOUNDINGS ANALYSIS: Processing versus Inversion [M]. Cecil and Ida Green Professor of Geophysics Stanford University,2004.
[6] Claerbout J. Earth soundings analysis [M]. Blackwell ScientificPubl,1992.
[7]库国正.道内插技术及其应用[J].石油地球物理勘探,1988,23(8):738-736.
[8] Spitz S. F-X域地震道内插[J].石油物探译丛,1992,(1):1-12.
[9]王振华.最小平方倾角分解与道间插值[J].石油地球物理勘探,1992,27(1):84-93.
[10]李国发. F-K域与F-X域联合实现道内插[J] .石油地球勘探物探,1995,30(5):693-701.
[11]国九英,周兴元. F-K域等道距道内插[J].石油地球物理勘探,1996,31(2):210-218.
[12]国九英,周兴元,俞寿朋. F-X域等道距道内插[J].石油地球物理勘探,1996,31(1):28-34.
[13]周兴元.F-X域地震道内插理论及其实现[J].石油地球物理勘探,1997,32(2):154-162.
[14]宜明理,严又生,魏新,吕德林,汪德雯. F-K域抗假频道内插[J].石油物探,2001,40(2):36-41.
[15]王建立,郭树祥,杨长春,张洪宙.广义谱分解地震道内插方法[J].石油地球物理勘探,2002,37(1):29-38.
[16]崔兴福,刘东奇,张关泉.小波变换实现地震道内插[J].石油地球物理勘探,2003,38:93-97.
[17]张红梅,刘洪.基于稀疏离散τ-p变换的叠后地震道内插[J].石油地球物理勘探,2006,41(3):281-285.
[18] Gulunay N.Seismic trace interpolation in the Fourier transform domain[J]. Geophysics ,2003,68(1):355-369.
[19] Larner K, Gibson B, Rothman D. Trace interpolation and?? design of seismic surveys [J]. Geophysics,1981,46, 407.
[20] Porsani J M. Seismic trace interpolation using half-step prediction filters[J] . Geophysics, 1999, 64, 1461-1467.
[21] Spitz S. Seismic trace interpolation in the f -x domain[J]. Geophysics, 1991, 56, 785-794.
[22] Vermeer G J O. Seismic wavefield sampling[C]: Soc Expl Geophys,1990.
[23] Claerbout J, Nichols D. Interpolation beyond aliasing by(t-x) domain PEFs[C]: 53rd Ann Internat Mtg ,Eur Assn Geosci Eng, Expanded Abstracts, paper A001, 1991.
[24] Lu L. Application of local slant-stack to trace interpolation[C]:55th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts,560-562,1985.
[25] Pieprzak A W, McClean J W. Trace interpolation of severely aliased events[C]: 58th Ann Internat Mtg, Soc Exp Geophys,Expanded Abstracts, 658-660,1988.
[26] Manin M, Spitz S. Wavefield de-aliasing for acquisition configurations leading to coarse sampling[C]: 65th Ann Internat Mtg Soc Exp, Geophys, Expanded Abstracts, 930-932,1995.
[27] Soubaras R. Spatial interpolation of aliased seismic data[C]: 67th Ann Internat Mtg, Soc Exp Geophys, Expanded Abstracts, 1167-1190, 1997.
[28] Spitz S. Trace interpolation beyond aliasing and without picking[C]:59th Ann Internat Mtg, Soc Exp Geophys, Expanded Abstracts,1121-1122,1989.
[29] Gulunay N, Chambers R E. Unaliased f -k domain trace interpolation (UFKI)[C]: 66th Ann Internat Mtg, Soc Expl Geophys ,Expanded Abstracts, 1461-1464,1996.
[30] Gulunay N,Chambers R E .Generalized f -k domain trace interpolation[C]: 67th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts, 1100-1103,1997.
[31] Guo J, Zhou X,Yang H J. Efficient f -k domainseismic trace interpolation for spatially aliased data[C]: 66th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts, 1457-1460,1996.
[32] Jakubowicz H. Wavefield reconstruction[C]: 64th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts, 1557-1560,1994.
[33] Ji J. Interpolation using prediction-error filter simulation(PEFs)[C]: 63rd Ann Internat Mtg, Soc Expl Geophys, ExpandedAbstracts, 1170-1173,1993.
[34] King G A, Leong T K,Flinchbaugh B E. The role of interpolation in seismic resolution[C]: 54th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstract葐?66-767,1984.
[2]黄德济,贺振华,包吉山.地震勘探资料数字处理[M].北京:地质出版社,1990.
[3]何樵登.地震勘探原理和方法[M].北京:石油大学出版社,1985.
[4]牟永光.地震勘探资料数字处理[M].北京:石油工业出版社,1981.
[5] Claerbout J. EARTH SOUNDINGS ANALYSIS: Processing versus Inversion [M]. Cecil and Ida Green Professor of Geophysics Stanford University,2004.
[6] Claerbout J. Earth soundings analysis [M]. Blackwell ScientificPubl,1992.
[7]库国正.道内插技术及其应用[J].石油地球物理勘探,1988,23(8):738-736.
[8] Spitz S. F-X域地震道内插[J].石油物探译丛,1992,(1):1-12.
[9]王振华.最小平方倾角分解与道间插值[J].石油地球物理勘探,1992,27(1):84-93.
[10]李国发. F-K域与F-X域联合实现道内插[J] .石油地球勘探物探,1995,30(5):693-701.
[11]国九英,周兴元. F-K域等道距道内插[J].石油地球物理勘探,1996,31(2):210-218.
[12]国九英,周兴元,俞寿朋. F-X域等道距道内插[J].石油地球物理勘探,1996,31(1):28-34.
[13]周兴元.F-X域地震道内插理论及其实现[J].石油地球物理勘探,1997,32(2):154-162.
[14]宜明理,严又生,魏新,吕德林,汪德雯. F-K域抗假频道内插[J].石油物探,2001,40(2):36-41.
[15]王建立,郭树祥,杨长春,张洪宙.广义谱分解地震道内插方法[J].石油地球物理勘探,2002,37(1):29-38.
[16]崔兴福,刘东奇,张关泉.小波变换实现地震道内插[J].石油地球物理勘探,2003,38:93-97.
[17]张红梅,刘洪.基于稀疏离散τ-p变换的叠后地震道内插[J].石油地球物理勘探,2006,41(3):281-285.
[18] Gulunay N.Seismic trace interpolation in the Fourier transform domain[J]. Geophysics ,2003,68(1):355-369.
[19] Larner K, Gibson B, Rothman D. Trace interpolation and?? design of seismic surveys [J]. Geophysics,1981,46, 407.
[20] Porsani J M. Seismic trace interpolation using half-step prediction filters[J] . Geophysics, 1999, 64, 1461-1467.
[21] Spitz S. Seismic trace interpolation in the f -x domain[J]. Geophysics, 1991, 56, 785-794.
[22] Vermeer G J O. Seismic wavefield sampling[C]: Soc Expl Geophys,1990.
[23] Claerbout J, Nichols D. Interpolation beyond aliasing by(t-x) domain PEFs[C]: 53rd Ann Internat Mtg ,Eur Assn Geosci Eng, Expanded Abstracts, paper A001, 1991.
[24] Lu L. Application of local slant-stack to trace interpolation[C]:55th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts,560-562,1985.
[25] Pieprzak A W, McClean J W. Trace interpolation of severely aliased events[C]: 58th Ann Internat Mtg, Soc Exp Geophys,Expanded Abstracts, 658-660,1988.
[26] Manin M, Spitz S. Wavefield de-aliasing for acquisition configurations leading to coarse sampling[C]: 65th Ann Internat Mtg Soc Exp, Geophys, Expanded Abstracts, 930-932,1995.
[27] Soubaras R. Spatial interpolation of aliased seismic data[C]: 67th Ann Internat Mtg, Soc Exp Geophys, Expanded Abstracts, 1167-1190, 1997.
[28] Spitz S. Trace interpolation beyond aliasing and without picking[C]:59th Ann Internat Mtg, Soc Exp Geophys, Expanded Abstracts,1121-1122,1989.
[29] Gulunay N, Chambers R E. Unaliased f -k domain trace interpolation (UFKI)[C]: 66th Ann Internat Mtg, Soc Expl Geophys ,Expanded Abstracts, 1461-1464,1996.
[30] Gulunay N,Chambers R E .Generalized f -k domain trace interpolation[C]: 67th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts, 1100-1103,1997.
[31] Guo J, Zhou X,Yang H J. Efficient f -k domainseismic trace interpolation for spatially aliased data[C]: 66th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts, 1457-1460,1996.
[32] Jakubowicz H. Wavefield reconstruction[C]: 64th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstracts, 1557-1560,1994.
[33] Ji J. Interpolation using prediction-error filter simulation(PEFs)[C]: 63rd Ann Internat Mtg, Soc Expl Geophys, ExpandedAbstracts, 1170-1173,1993.
[34] King G A, Leong T K,Flinchbaugh B E. The role of interpolation in seismic resolution[C]: 54th Ann Internat Mtg, Soc Expl Geophys, Expanded Abstract葐?66-767,1984.