地震勘探数据压缩相关技术研究及应用
|
摘要
随着多道多分量地震勘探信息采集技术的实施，获取勘探的数据量以惊人的速度增长，给存储和传输都带来了一定的困难，所以对于数据压缩的要求越来越迫切，成为地球物理领域中重要的研究内容。目前国内外地震勘探数据压缩技术的研究相对比较薄弱，还需要进一步发展和完善，使这方面的研究工作更具有系统性、综合性、实用性。本次选择了“地震勘探数据压缩相关技术研究及应用”为研究内容，并选择了两种技术重点研究，即离散余弦变换编码和正交小波变换编码。在离散余弦变换编码中，研究的重点是离散余弦变换的去相关性和对数据的改造重组能力；利用正交小波变换进行小波分频特性分析处理，得到了较好的压缩结果。这两部分都编制了有实用价值的软件。
离散余弦变换编码中用到的主要技术是离散余弦变换和游程编码。虽然离散余弦变换已经广泛的应用到信号和图象处理中，但在地震数据上的应用的研究还很少，因为地震数据和自然图象的数据除了相似之处还有差别——地震数据动态范围大、中低频信息丰富。论文中，首先把离散余弦变换编码应用到地震模型中，总结了规律后，再应用到地震勘探的实际数据处理中，通过实验和分析知道该方法有这样几个特点：
离散余弦变换能够较好的利用地震数据道与道之间的相关性，经过该方法处理后变换系数的能量集中，有利于合理少分配比特数。
变换系数用游程编码进行量化时，对高频信息不灵敏，使该方法的应用受到了很多限制。
虽然离散余弦变换是正交变换，但是地震数据经过正变换和反变换后数据还是有细微的变化，但这不影响地质解释，因为反映在地震剖面上这点儿变化是肉眼识别不出来的。
压缩处理后，不同压缩比时恢复的地震数据的畸变程度不同，并伴随假高频现象出现。
该方法还有一个缺点，那就是会出现“方块”效应，这也影响该方法的应用。经过分析论证，认为解决的最好办法就是降低压缩比，这样“方块”效应不明显，肉眼看不出来。
利用离散余弦变换编码进行地震数据压缩时，其效果与压缩的地震数据的松散度、信噪比等有一定的关系，很难给出定量的说明。需要根据数据的具体情况和地质解释精度要求，来给出合适的压缩比。
由于离散余弦变换编码的压缩程度比较小，而且效果不是很理想，所以我们还采用小波变换理论来提高数据的压缩比。小波变换被誉为“数学显微镜”，是一个时间和频率的局域变换，能够通过伸缩和平移对函数和信号进行多尺度细化分析。我们应用这一特性把地震数据进行小波变换，把地震数据分成低频低波数、低频高波数、高频低波数和高频高波数四个部分，根据数据特性的差别进行编码处理。处理之前我们应该弄清楚以下几个问题：第一，对数据反映的成层性进行分析，了解其频率和波数的特征。第二，不同数据应选择不同小波基，因为不同的小波基效果是不同的。本次研究采用双正交的小波，以保证变换是可逆的，并没有相位失真。本文用的是 daubechies小波。第三，研究清楚已知数据的分辨率，根据要求选择适当的压缩比。
为了进一步提高地震数据的压缩比，我们把小波变换编码和离散余弦变换编码联合起来应用。通过两个实际地震数据的处理，并经过详细的实验分析，认识到以下几个特点：
根据多尺度分析的细化，在频率域和空间域上把地震数据分成不同的块体，然后根据重要性的不同，加以处理。
压缩比比较大，而且能够达到比较满意的效果。
没有“方块”效应。
通过分析，小波编码对于不同的地震数据压缩都有效果，但是压缩比不尽相同，恢复出来的地震剖面的效果差异也很大。因此，在压缩地震数据的时候，应该根据实际情况调整压缩比，以保证恢复出来的地震剖面能够满足需要。如果对地震剖面要求很高，也就是不丢失层间信息的细节结构，那么我们推荐压缩比应在10以下。
通过研究，发现地震数据的压缩处理不同于其它的图象压缩，不能用确定的标准和固定的参数进行压缩，应根据实际情况进行实验确定压缩比。这样才能达到既节省存储空间，又加快传输速度，最终实现地震数据压缩的目的。
When the technique of collection develops that it is multi-track and multi-variable, data of protecting are increasing fast, and these give us matter for memory and transmission, so it is in urgent need of data compression, and it is going to be the important content of research in the geophysical domain. At present the weak research is in data compression of seismic prospecting, and we need to develop and perfect this technique in order to make it systematic, all-around, practical. I choose “the correlative technical research and application of data compression in seismic protecting” as my content of study, and emphasize two techniques: the discrete cosine transform coding and the orthogonal wavelet transformation coding. In the discrete cosine transform coding, the keystone is that discrete cosine transform coding can wipe off the correlation and recompose data. In the orthogonal wavelet transformation coding, it is to make use of wavelet frequency splitting that may deal with data, and then get the better result of compression. The applied software of the two parts has been executed.
The discrete cosine transform coding consists of discrete cosine transform and
run length coding mainly. Although discrete cosine transform has been applied to signal and image processing generally, it is seldom applied to data processing of protecting, for there are much difference between them, which seismic prospecting data are of big dynamic range and wealth of low-intermediate frequency information. In the paper, discrete cosine transform is first applied to seismic model. After we make a summary, it is applied to real data of seismic protecting by us. By the way of experiment we find this way has several characters:
Discrete cosine transform can utilize correlativity of traces, and concentrate energy of data, so it may distribute bit number reasonably.
When transformation coefficient is quantized by run length coding, it is not sensitive to high frequency, therefore this way is restricted within narrow limits
Discrete cosine transform is orthogonal transform, but there are some small changes between real data and transpositional data. For these changes are difficult to identify on the seismic section, it may not affect geologic interpretation.
After compression disposing, along with the different ratio of compression there is different to aberrance in the resumptive data with the appearance of false high frequency.
This way has one shortcoming, which can appear the domino offect of “diamonds”, so restricted the application of the method. It is to reduce the compression ratio that is the best resoluble measure, and the domino offect of “diamonds” can not be identified by eyes.
It is difficult to show quantificationally that data of protecting is compressed by discrete cosine transform, as its effect is relative with the incompact degree and Signal-to-Noise. Hence we give one appropriate Signal-to-Noise on basis of practical situation.
Signal-to-Noise is small utilizing discrete cosine transform, and its effect is not perfect, so we need increase signal-to noise with wavelet transformation. Wavelet transformation is called “mathematical microscope”, and it is a local transform in the time-frequency domain, then it may analyze function and signal on
multi-resolution by flex and translation. We process data using wavelet transform on above character, and can divide data into four parts: low frequency and low wave number, low frequency and high wave number, high frequency and low wave number, high frequency and high wave number. And then we process data to code on basis of character. We must understand lower several questions: firstly, we ought to analyze the reason of bedding in order to know data’s character of frequency and wave number. Secondly, different wavelet basis has different effect. The research adopts double quadrature wavelet, which can make the transform reversible and avoid the phasic distortion. Here is daubechies wavelet. Thirdly, we must know the data’s resolution, before we c
离散余弦变换编码中用到的主要技术是离散余弦变换和游程编码。虽然离散余弦变换已经广泛的应用到信号和图象处理中，但在地震数据上的应用的研究还很少，因为地震数据和自然图象的数据除了相似之处还有差别——地震数据动态范围大、中低频信息丰富。论文中，首先把离散余弦变换编码应用到地震模型中，总结了规律后，再应用到地震勘探的实际数据处理中，通过实验和分析知道该方法有这样几个特点：
离散余弦变换能够较好的利用地震数据道与道之间的相关性，经过该方法处理后变换系数的能量集中，有利于合理少分配比特数。
变换系数用游程编码进行量化时，对高频信息不灵敏，使该方法的应用受到了很多限制。
虽然离散余弦变换是正交变换，但是地震数据经过正变换和反变换后数据还是有细微的变化，但这不影响地质解释，因为反映在地震剖面上这点儿变化是肉眼识别不出来的。
压缩处理后，不同压缩比时恢复的地震数据的畸变程度不同，并伴随假高频现象出现。
该方法还有一个缺点，那就是会出现“方块”效应，这也影响该方法的应用。经过分析论证，认为解决的最好办法就是降低压缩比，这样“方块”效应不明显，肉眼看不出来。
利用离散余弦变换编码进行地震数据压缩时，其效果与压缩的地震数据的松散度、信噪比等有一定的关系，很难给出定量的说明。需要根据数据的具体情况和地质解释精度要求，来给出合适的压缩比。
由于离散余弦变换编码的压缩程度比较小，而且效果不是很理想，所以我们还采用小波变换理论来提高数据的压缩比。小波变换被誉为“数学显微镜”，是一个时间和频率的局域变换，能够通过伸缩和平移对函数和信号进行多尺度细化分析。我们应用这一特性把地震数据进行小波变换，把地震数据分成低频低波数、低频高波数、高频低波数和高频高波数四个部分，根据数据特性的差别进行编码处理。处理之前我们应该弄清楚以下几个问题：第一，对数据反映的成层性进行分析，了解其频率和波数的特征。第二，不同数据应选择不同小波基，因为不同的小波基效果是不同的。本次研究采用双正交的小波，以保证变换是可逆的，并没有相位失真。本文用的是 daubechies小波。第三，研究清楚已知数据的分辨率，根据要求选择适当的压缩比。
为了进一步提高地震数据的压缩比，我们把小波变换编码和离散余弦变换编码联合起来应用。通过两个实际地震数据的处理，并经过详细的实验分析，认识到以下几个特点：
根据多尺度分析的细化，在频率域和空间域上把地震数据分成不同的块体，然后根据重要性的不同，加以处理。
压缩比比较大，而且能够达到比较满意的效果。
没有“方块”效应。
通过分析，小波编码对于不同的地震数据压缩都有效果，但是压缩比不尽相同，恢复出来的地震剖面的效果差异也很大。因此，在压缩地震数据的时候，应该根据实际情况调整压缩比，以保证恢复出来的地震剖面能够满足需要。如果对地震剖面要求很高，也就是不丢失层间信息的细节结构，那么我们推荐压缩比应在10以下。
通过研究，发现地震数据的压缩处理不同于其它的图象压缩，不能用确定的标准和固定的参数进行压缩，应根据实际情况进行实验确定压缩比。这样才能达到既节省存储空间，又加快传输速度，最终实现地震数据压缩的目的。
When the technique of collection develops that it is multi-track and multi-variable, data of protecting are increasing fast, and these give us matter for memory and transmission, so it is in urgent need of data compression, and it is going to be the important content of research in the geophysical domain. At present the weak research is in data compression of seismic prospecting, and we need to develop and perfect this technique in order to make it systematic, all-around, practical. I choose “the correlative technical research and application of data compression in seismic protecting” as my content of study, and emphasize two techniques: the discrete cosine transform coding and the orthogonal wavelet transformation coding. In the discrete cosine transform coding, the keystone is that discrete cosine transform coding can wipe off the correlation and recompose data. In the orthogonal wavelet transformation coding, it is to make use of wavelet frequency splitting that may deal with data, and then get the better result of compression. The applied software of the two parts has been executed.
The discrete cosine transform coding consists of discrete cosine transform and
run length coding mainly. Although discrete cosine transform has been applied to signal and image processing generally, it is seldom applied to data processing of protecting, for there are much difference between them, which seismic prospecting data are of big dynamic range and wealth of low-intermediate frequency information. In the paper, discrete cosine transform is first applied to seismic model. After we make a summary, it is applied to real data of seismic protecting by us. By the way of experiment we find this way has several characters:
Discrete cosine transform can utilize correlativity of traces, and concentrate energy of data, so it may distribute bit number reasonably.
When transformation coefficient is quantized by run length coding, it is not sensitive to high frequency, therefore this way is restricted within narrow limits
Discrete cosine transform is orthogonal transform, but there are some small changes between real data and transpositional data. For these changes are difficult to identify on the seismic section, it may not affect geologic interpretation.
After compression disposing, along with the different ratio of compression there is different to aberrance in the resumptive data with the appearance of false high frequency.
This way has one shortcoming, which can appear the domino offect of “diamonds”, so restricted the application of the method. It is to reduce the compression ratio that is the best resoluble measure, and the domino offect of “diamonds” can not be identified by eyes.
It is difficult to show quantificationally that data of protecting is compressed by discrete cosine transform, as its effect is relative with the incompact degree and Signal-to-Noise. Hence we give one appropriate Signal-to-Noise on basis of practical situation.
Signal-to-Noise is small utilizing discrete cosine transform, and its effect is not perfect, so we need increase signal-to noise with wavelet transformation. Wavelet transformation is called “mathematical microscope”, and it is a local transform in the time-frequency domain, then it may analyze function and signal on
multi-resolution by flex and translation. We process data using wavelet transform on above character, and can divide data into four parts: low frequency and low wave number, low frequency and high wave number, high frequency and low wave number, high frequency and high wave number. And then we process data to code on basis of character. We must understand lower several questions: firstly, we ought to analyze the reason of bedding in order to know data’s character of frequency and wave number. Secondly, different wavelet basis has different effect. The research adopts double quadrature wavelet, which can make the transform reversible and avoid the phasic distortion. Here is daubechies wavelet. Thirdly, we must know the data’s resolution, before we c
引文
[1] 程乾生. 信号数字处理的数学原理. 北京: 石油工业出版社，1979.
[2] 托马斯J.林奇. 数据压缩技术及其应用. 吴家安, 杜淑玲 译. 北京: 人民邮电出版社，1989.
[3] 刘贵忠, 邸双亮. 小波分析及其应用. 西安: 西安电子科技大学出版社, 1992.
[4] 秦前清, 杨宗凯. 实用小波分析. 西安: 西安电子科技大学出版社, 1994.
[5] 吴乐南. 数据压缩的原理与应用. 北京: 电子出版社, 1995.
[6] 袁玫, 袁文. 数据压缩技术及应用. 北京: 电子工业出版社, 1995.
[7] 杨福生. 小波变换的工程分析与应用. 北京: 科学出版社, 2001.
[8] 孙圣和, 陆哲明. 矢量量化技术及应用. 北京: 科学出版社, 2002.
[9] 余品能. 二维离散余弦变换的FFT及FPT混合算法. 石油地球物理勘探, 1994, 29 (4): 468~473.
[10] 赵改善. 地震数据通用压缩文件格式设计. 石油物探, 1994, 33 (3): 20~25.
[11] 赵改善. 小波变换与地震数据压缩. 石油物探, 1994,33(4), 19~28.
[12] 王真理等. 利用小波变换压缩地震勘探数据. 石油地球物理勘探, 1995, 30(1), 86~97.
[13] 吴乐南. 声波勘探信号的预测编码. 石油地球物理勘探, 1995, 30 (4): 505~508.
[14] 孟昭波, 杨丽华. 地震资料的小波压缩. 石油地球物理勘探, 1995, 30(增): 70~75.
[15] 许刚等. 小波理论在图象中的应用. 武汉水利电力大学学报, 1996, 29(6).
[16] 智西湖. 小波变换与图象数据压缩. 河南大学学报（自然科学版）, 1996, 26(2), 32~36.
[17] 洪海等. 小波变换在静态图像数据压缩中的应用. 小型微型计算机系统, 1996, 17(11), 1~5.
[19] 刘琦等. 基于二维小波分解的图象编码. 计算机辅助设计与图形学学报, 1996, 8(1), 18~25.
[20] 薛晓辉等. 基于小波变换的静止图象压缩系统. 宇航学报, 1997, 18(1).
[21] 张学工, 李衍达, 盛硕. 几种现代信息处理技术和计算机网络在勘探地球物理中的
应用. 地球物理学报, 1997, 40 (增): 275~291.
[22] 陈东义等. Mallat塔式算法在一维信号处理中的应用研究. 重庆大学学报（自然科学报）, 1999, 22(4).
[23] 唐向宏等. 利用二维M带小波变换进行地震数据压缩. 石油地球物理勘探, 1999, 34(2).
[24] 郝俊瑞等. 图象压缩中小波基的选择. 桂林电子工业学院学报, 2000, 20(2).
[25] 陈雷霆等. 小波图像编码与JPEG图像编码的比较研究. 计算机应用, 2000, 20(增).
[26] 冯占林等. 基于小波变换的地震勘探数据压缩的工程分析. 清华大学学报（自然科学版）, 2001, 41(4/5).
[27] 刘财, 王培茂等. 离散余弦变换(DCT)编码在地震勘探数据压缩上的应用. 吉林大学学报（地球科学版）, 2004, 34(2), 277~282.
[28] 王卫国等. 嵌入式零树小波编码及其改进算法的研究.
[29] WEN-HSIUNG, C. HARRISON SMITH, AND S. C. FRALICK. A fast computational algorithm for the discrete cosine transform. IEEE. Transaction On Communications, 1977, 25(9).
[30] Mallat S. A theory multiresolution signal decomposition: The wavelet representation. IEEE Trans. Pattern Anal. Mach. Intell, 1989, 11, 674~693.
[31] Mallat S. Multifrequency channel decompositions of images and wavelet model. IEEE Trans. On ASSP, 1989, 37(12), 2091~2110.
[32] Daubechies I. Orthonormal bases of compactly supported wavelets. IEEE Trans. Inform. Theory, 1990, 36, 961~1005.
[33] Andreas S. Spanias, Stefan B. Jonsson, Samuel D. Stearns. Transform methods for seismic data compression[J]. IEEE Transaction On Geoscience And Remote Sensing, 1991, 29(3): 407~415.
[34] Steffen P, Heller P N, Gopinuth R A. Theory of images and wavelet models. IEEE Trans on SP, 1993, 41(12), 3497~3511.
[35] J. M. Shapiro. Embedded image coding using zerotrees of wavelet coefficients. IEEE Trans. On Signal Processing, 1993, 41(12), 3445~3462.
[36] A. Zandi, J. Allen, E. Schwartz, and M. Boliek, CREW: Compression with reversible
embedded wavelets, IEEE Data Compression Conference, Snowbird, UT, 1995, 3, 212~221.
[37] A. Said and W. A. Perlman. An image multiresolution representation for lossless and lossy compression. IEEE Trans. On Image Processing, 1996, 5(9), 1303~1310.
[38] A. Said and A. Perlman. A new fast and efficient image codec based on set partitioning in hierarchical trees. IEEE Transactions on Circuits and Systems for Video Technology, 1996, 6(3), 245~250.
[39] Joonas Lehtinen. Limiting distortion of a wavelet image codec. Acta Cybernetica, 1999, 14, 1~16.
[40] Robi Polikar. Wavelettutorial.
http://engineering.rowan.edu/~polikar/WAVELETS/Wtutorial.html
[41] http://guohanwei.51.net/
[2] 托马斯J.林奇. 数据压缩技术及其应用. 吴家安, 杜淑玲 译. 北京: 人民邮电出版社，1989.
[3] 刘贵忠, 邸双亮. 小波分析及其应用. 西安: 西安电子科技大学出版社, 1992.
[4] 秦前清, 杨宗凯. 实用小波分析. 西安: 西安电子科技大学出版社, 1994.
[5] 吴乐南. 数据压缩的原理与应用. 北京: 电子出版社, 1995.
[6] 袁玫, 袁文. 数据压缩技术及应用. 北京: 电子工业出版社, 1995.
[7] 杨福生. 小波变换的工程分析与应用. 北京: 科学出版社, 2001.
[8] 孙圣和, 陆哲明. 矢量量化技术及应用. 北京: 科学出版社, 2002.
[9] 余品能. 二维离散余弦变换的FFT及FPT混合算法. 石油地球物理勘探, 1994, 29 (4): 468~473.
[10] 赵改善. 地震数据通用压缩文件格式设计. 石油物探, 1994, 33 (3): 20~25.
[11] 赵改善. 小波变换与地震数据压缩. 石油物探, 1994,33(4), 19~28.
[12] 王真理等. 利用小波变换压缩地震勘探数据. 石油地球物理勘探, 1995, 30(1), 86~97.
[13] 吴乐南. 声波勘探信号的预测编码. 石油地球物理勘探, 1995, 30 (4): 505~508.
[14] 孟昭波, 杨丽华. 地震资料的小波压缩. 石油地球物理勘探, 1995, 30(增): 70~75.
[15] 许刚等. 小波理论在图象中的应用. 武汉水利电力大学学报, 1996, 29(6).
[16] 智西湖. 小波变换与图象数据压缩. 河南大学学报（自然科学版）, 1996, 26(2), 32~36.
[17] 洪海等. 小波变换在静态图像数据压缩中的应用. 小型微型计算机系统, 1996, 17(11), 1~5.
[19] 刘琦等. 基于二维小波分解的图象编码. 计算机辅助设计与图形学学报, 1996, 8(1), 18~25.
[20] 薛晓辉等. 基于小波变换的静止图象压缩系统. 宇航学报, 1997, 18(1).
[21] 张学工, 李衍达, 盛硕. 几种现代信息处理技术和计算机网络在勘探地球物理中的
应用. 地球物理学报, 1997, 40 (增): 275~291.
[22] 陈东义等. Mallat塔式算法在一维信号处理中的应用研究. 重庆大学学报（自然科学报）, 1999, 22(4).
[23] 唐向宏等. 利用二维M带小波变换进行地震数据压缩. 石油地球物理勘探, 1999, 34(2).
[24] 郝俊瑞等. 图象压缩中小波基的选择. 桂林电子工业学院学报, 2000, 20(2).
[25] 陈雷霆等. 小波图像编码与JPEG图像编码的比较研究. 计算机应用, 2000, 20(增).
[26] 冯占林等. 基于小波变换的地震勘探数据压缩的工程分析. 清华大学学报（自然科学版）, 2001, 41(4/5).
[27] 刘财, 王培茂等. 离散余弦变换(DCT)编码在地震勘探数据压缩上的应用. 吉林大学学报（地球科学版）, 2004, 34(2), 277~282.
[28] 王卫国等. 嵌入式零树小波编码及其改进算法的研究.
[29] WEN-HSIUNG, C. HARRISON SMITH, AND S. C. FRALICK. A fast computational algorithm for the discrete cosine transform. IEEE. Transaction On Communications, 1977, 25(9).
[30] Mallat S. A theory multiresolution signal decomposition: The wavelet representation. IEEE Trans. Pattern Anal. Mach. Intell, 1989, 11, 674~693.
[31] Mallat S. Multifrequency channel decompositions of images and wavelet model. IEEE Trans. On ASSP, 1989, 37(12), 2091~2110.
[32] Daubechies I. Orthonormal bases of compactly supported wavelets. IEEE Trans. Inform. Theory, 1990, 36, 961~1005.
[33] Andreas S. Spanias, Stefan B. Jonsson, Samuel D. Stearns. Transform methods for seismic data compression[J]. IEEE Transaction On Geoscience And Remote Sensing, 1991, 29(3): 407~415.
[34] Steffen P, Heller P N, Gopinuth R A. Theory of images and wavelet models. IEEE Trans on SP, 1993, 41(12), 3497~3511.
[35] J. M. Shapiro. Embedded image coding using zerotrees of wavelet coefficients. IEEE Trans. On Signal Processing, 1993, 41(12), 3445~3462.
[36] A. Zandi, J. Allen, E. Schwartz, and M. Boliek, CREW: Compression with reversible
embedded wavelets, IEEE Data Compression Conference, Snowbird, UT, 1995, 3, 212~221.
[37] A. Said and W. A. Perlman. An image multiresolution representation for lossless and lossy compression. IEEE Trans. On Image Processing, 1996, 5(9), 1303~1310.
[38] A. Said and A. Perlman. A new fast and efficient image codec based on set partitioning in hierarchical trees. IEEE Transactions on Circuits and Systems for Video Technology, 1996, 6(3), 245~250.
[39] Joonas Lehtinen. Limiting distortion of a wavelet image codec. Acta Cybernetica, 1999, 14, 1~16.
[40] Robi Polikar. Wavelettutorial.
http://engineering.rowan.edu/~polikar/WAVELETS/Wtutorial.html
[41] http://guohanwei.51.net/
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |