论文部分内容阅读
由于地震子波的时变性，在假设子波为非时变的情况下，用脉冲反褶积对整个地震道进行处理，所得地震剖面会出现浅层频率高，深层频率低的现象。实践证明，采用开时窗的方法在时窗短时，这种反褶积的效果也很差。Burg反褶积虽然采用滑动时窗，可适应子波的时变性，但由于受Levinson关系的约束，仍然摆脱不了Toeplitz矩阵的性质，不能完全克服脉冲反褶积的缺限。为此，我们提出了无约束正反向预测误差最小反褶积方法。该方法无需解Toeplitz矩阵，也不要求子波是非时变的，并采用Marple递推算法快速实现。理论分析和实际应用的效果表明，该方法是一种对小时窗和大时窗均有效且稳健的反褶积方法。
Due to the time-varying seismic wavelet, under the assumption of time-invariant wavelets, pulse deconvolution is used to process the entire seismic trace. The resulting seismic profile will have shallow shallow frequencies and low deep frequencies. Practice has proved that the use of open windows method in the short time window, the effect of this deconvolution is poor. Although Burg deconvolution can adapt to the time-varying of wavelet by sliding window, it still can not get rid of the Toeplitz matrix because of the Levinson relation, and can not completely overcome the shortcoming of deconvolution. To this end, we propose the least deconvolution method for unconstrained forward and backward prediction errors. The method does not need to solve the Toeplitz matrix and does not require the wavelet to be time-invariant and can be quickly implemented using Marple recursion algorithm. Theoretical analysis and practical application results show that this method is an effective and robust deconvolution method for small time windows and large time windows.