关键词: 数据重建, 流预测滤波器, 非平稳地震数据, 虚拟一次波

Abstract: Seismic data is essentially nonstationary. How to solve the problem of missing data interpolation problem for complex nonstationary seismic wave fields is one of the keys in data processing. Prediction filters play an important role in the seismic data processing and analysis. The technology can solve the data missing problem efficiently; however, the traditional stationary prediction filter cannot fit the nonstationary feature of seismic data, so to develop an adaptive prediction interpolation method for complex seismic wave fields has an important industrial value. In the paper, we introduced the conception of "Streaming" into the prediction filter, the filter can update its coefficients incrementally by accepting one new data point at one time. This process only needs to compute the vector dot product, which improves the operation efficiency and reduces the memory space. In addition, we can do the inverse operation for seismic data by using the streaming prediction filter, which can reconstruct the missing data quickly and effectively. By using the multiple kinetic information to construct the pseudo-primaries through cross-correlation technology, we effectively solved the problem of inaccurate estimation of the filter coefficients in the missing data position, provided more reasonable filter estimates for the interpolation process, and reconstructed the nonstationary seismic data. that the filter estimation is inaccurate in the position of missing data. The pseudoprimaries provide reasonable filter estimation in the interpolation and can help solving the problem that the reconstruction of nonstationary seismic data better. The test results of the Sigsbee2B model and the field data show that the new method can reasonably reconstruct the missing data for complex seismic information.

Key words: data reconstruction, streaming prediction filter, nonstationary seismic data, pseudoprimary