基于傅里叶变换的地震数据规则化和插值

发布时间：2018-05-02 06:25

本文选题：地震数据 + 不规则采样　；参考：《浙江大学》2015年硕士论文

【摘要】：在地球物理地震勘探的实际数据采集中,由于工区地表环境(如河流,湖泊,山谷等)、地表障碍物(道路,桥梁等)以及禁采区的影响,检波器的布置常常是不规则的。另外由于检波器与地表耦合效果不好,环境干扰或者仪器自身的原因会产生一些废道和废炮,这些无用的地震道需要在处理过程中剔除。在海上地震勘探中,拖缆的羽状漂移也使得其偏离所设计的规则网格位置。随着地震勘探的不断深入,采集到的数据量也越来越大,一方面我们希望得到密集采样的数据以获得更准确的地下结构,另一方面考虑到勘探成本,需要在空间方向进行稀疏采样。因此我们得到的数据常常是稀疏和不规则的。在后续的数据处理过程中,由于稀疏和不规则采样会产生噪声,并且会影响一些多道处理技术,如叠加、波动方程偏移、多次波滤除和时移地震等,因此对地震数据进行重建是非常关键和必要的。本文的研究内容包括地震数据规则化和地震数据插值,都可称为地震数据重建。但是前者处理后的数据的位置和实际采样点的位置完全不同,相当于在此工区重新设计了一套观测系统,重建后地震道数可以等于原始不规则采样的地震道数,也可以大于或小于不规则地震数据的道数。后者处理后的数据仍然包含实际资料所有的采样位置和采集到的数据,只是对空缺地震道进行插值,重建后的地震道数一定大于非规则采样的地震道数。在地震数据规则化处理中,研究了反泄露傅里叶变换,提出一种改进措施,将其中非均匀离散傅里叶变换公式经过简单处理后,使用非均匀快速傅里叶变换,大大提高了计算效率。模型试验证明了结合NFFT的反泄露傅里叶变换在规则化处理中的有效性。在地震数据插值中,研究了凸集投影法和迭代阈值法,对迭代阈值法提出改进措施,即结合线性插值的迭代阈值法。模型试验证明了本文所提出的方法比凸集投影法和迭代阈值法插值效果好。本文地震数据规则化和插值处理都是在FX域对部分频率成分进行的,与TX域对每一个时间成分进行重建相比,大大减小了计算量,并且可提高重建准确度。
[Abstract]:In the actual data acquisition of geophysical seismic exploration, the geophone arrangement is often irregular due to the influence of the surface environment (such as rivers, lakes, valleys, etc.), surface obstacles (roads, bridges, etc.) and no mining areas in the work area. In addition due to the poor coupling effect between geophone and surface environmental interference or the instrument itself will produce some abandoned tracks and guns which need to be eliminated in the process of processing. In offshore seismic exploration, the plume drift of towing cable also makes it deviate from the designed regular grid position. With the deepening of seismic exploration, the amount of data collected is becoming larger and larger. On the one hand, we want to obtain dense sampling data in order to obtain more accurate underground structure, on the other hand, considering the exploration cost, Sparse sampling is required in the spatial direction. So the data we get are often sparse and irregular. In the subsequent data processing process, the sparse and irregular sampling will produce noise, and will affect some multi-channel processing techniques, such as stack, wave equation migration, multiple wave filtering and time-lapse earthquakes, etc. Therefore, the reconstruction of seismic data is very important and necessary. The contents of this paper include regularization of seismic data and interpolation of seismic data, which can be called seismic data reconstruction. However, the position of the data processed by the former is completely different from that of the actual sampling points, which is equivalent to the redesign of a set of observation systems in this work area. After reconstruction, the number of seismic traces can be equal to the number of seismic traces of the original irregular sampling. It can also be greater than or less than the trace number of irregular seismic data. The processed data of the latter still contain all the sampling locations and collected data of the actual data, but only interpolate the vacant seismic traces, and the reconstructed seismic traces must be larger than the irregular sampling seismic traces. In the regularization of seismic data, the anti-leakage Fourier transform is studied, and an improved method is put forward, in which the non-uniform discrete Fourier transform formula is simply processed and the non-uniform fast Fourier transform is used. The calculation efficiency is greatly improved. The model tests show the effectiveness of the anti-leakage Fourier transform combined with NFFT in regularization. In the interpolation of seismic data, the projection method of convex set and the iterative threshold method are studied, and the improvement measures of iterative threshold method are put forward, that is, the iterative threshold method combined with linear interpolation. The model tests show that the proposed method is more effective than convex set projection method and iterative threshold method. In this paper, the regularization and interpolation of seismic data are carried out in FX domain. Compared with the reconstruction of each time component in TX domain, the computation is greatly reduced and the reconstruction accuracy is improved.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：P631.44

【参考文献】