Burg 算法在地震数据反褶积中的应用

The application of Burg algorithm in seismic data deconvolution

Burg 算法的一个重要概念是在求自相关函数时,对数据时窗外仅作最大随机性假设,而不认为等于零。用这一概念,可以外推自相关函数。当用 Burg算法求反滤波器预测误差因子 a_i^((m))时,既不计算自相关函数,也不计算反滤波因子的全部系数值,而只需要计算因子最末一点的系数 a_m^((m)),并且这一计算仅限定在数据时窗范围内进行。本文将 Burg 算法具体应用于地震数据反褶积中,在求 a_m^((m))时作了部分的修改,提高了稳定程度,改善了反褶积的地震剖面质量。从已进行的实际地震数据试算结果来看,只要处理参数使用恰当,地震剖面上信号的分辨率均有十分明显的提高,并且信噪比也有所改善。

An essential point of Burg algorithm is that the outside of a time window is only assumed as the maximum randomness instead of zero when autocorrelation function is calculated.This idea can be used to make extrapolation of autocorrelation function.Using Burg algorithm to estimate prediction error factor a^(m)_i of an inverse fil- ter,we only calculate the last coefficient a^(m)_m of the inverse filtering factor,neither the autocorrelation function nor all coefficients of the factor;what is more,the calculation is only restricted within a time window. Burg algorithm is applied to seismic data deconvolution,When used in calculating a^(m)_(m),the algorithm is partly revised,so that we improve deconvolution stability and seismic section quality,Having analysed the seismic data processed with this algorithm,we can find that both the resolution and the S/N ratio of the seismic section are improved on condition that processing parameters are appropriate.