论文部分内容阅读
地震数据本质上是非平稳的，如何解决复杂非平稳地震波场的数据缺失问题是地震勘探数据处理的重要环节之一。预测滤波器在地震数据处理和分析中起到重要的作用，该技术可以有效地解决地震数据缺失问题，但传统的平稳预测滤波方法无法很好地适应地震数据的非平稳特征，因此开发高效的复杂地震波场自适应预测插值方法具有重要的工业价值。本文将预测滤波器加入“流处理”的概念，滤波器系数随着地震数据的变化而同时更新，此计算过程仅需矢量点积运算，能够提高计算效率并降低内存空间，以此为基础开发基于流预测滤波的地震数据插值方法。利用多次波的动力学信息，通过互相关技术构建虚拟一次波，有效地解决在缺失数据位置滤波系数估计不准的问题，为插值过程提供更为合理的滤波器估计，更好地解决非平稳地震数据的重建问题。对Sigsbee 2B模型和实际数据的测试结果表明，该方法可以合理地针对复杂地震信息完成缺失数据的重建。
Seismic data are non-stationary in nature. How to solve the data missing problem of complex non-stationary seismic wavefields is one of the most important aspects of seismic data processing. Predictive filter plays an important role in seismic data processing and analysis. This technique can effectively solve the problem of missing seismic data. However, the traditional stationary prediction filtering method can not well adapt to the non-stationary characteristics of seismic data. Therefore, the development of efficient Complex seismic wavefield adaptive prediction interpolation method has important industrial value. In this paper, the concept of “predictive filter” is added to “stream processing”. The filter coefficients are updated simultaneously with the change of seismic data. The calculation process only needs vector dot product operation, which can improve the computational efficiency and reduce the memory space. Based on this, Seismic Data Interpolation Method Based on Flow Prediction Filtering. By using the kinetic information of multiple waves and constructing the virtual primary wave through the cross-correlation technique, the problem of inaccurate estimation of the filter coefficient at the missing data location is effectively solved, and a more reasonable filter estimation is provided for the interpolation process so as to better solve the problem of non- Reconstruction of stable seismic data. The test results of Sigsbee 2B model and real data show that this method can reasonably complete the reconstruction of missing data for complex seismic information.