基于稀疏约束贝叶斯估计的相对波阻抗反演

Relative acoustic impedance inversion based on Cauchy sparseness constraint Bayesian estimation.

1.中国地震局地球物理研究所,北京100081;2.北京大学地球物理学系,北京100871;3.中国石油化工股份有限公司石油勘探开发研究院,北京100083

Di Haibin, Institute of Geophysics,China Earthquake Administration,Beijing 100081,China

地震道积分是一种利用地震资料进行的无约束反演技术,可以方便地得到地层的相对波阻抗，但采用稀疏反演的反射系数递推反演相对波阻抗,横向连续性差、可用性低。给出了一种基于稀疏约束贝叶斯估计的地震相对波阻抗反演算法,即在反射系数Cauchy概率分布稀疏约束下,基于贝叶斯最大后验概率估计,采用预条件共轭梯度法估计地震子波和反射系数,从而得到高分辨率的反褶积结果,进而得到高分辨率的相对波阻抗剖面。利用理论模型和实际数据对算法进行了验证,结果表明:基于稀疏约束贝叶斯估计的相对波阻抗反演方法可行;与直接法稀疏反演相比,预条件共轭梯度法稀疏反演精度高、收敛快、数值计算稳定。

Seismic trace integration is a nonconstraint inversion technique.However,the relative acoustics impedance recursed by reflectivity obtained from sparse inversion has bad lateral continuity and worse adaptability.Based on Cauchy sparseness constraint Bayesian estimation,an inversion algorithm for seismic relative acoustic impedance was proposed.Assuming the Cauchy sparseness distribution of reflectivity,based on Bayesian maximum posterior probability estimation,precondition conjugate gradient inversion algorithm was adopted to realize the simultaneous estimation of reflectivity and wavelet.Then highresolution deconvolution result was obtained,and highresolution relative acoustic impedance profile was achieved.The inversion algorithm was tested by both theoretical model and real 2D seismic data.Test results indicate that,relative acoustic impedance inversion based on Cauchy sparseness constraint Bayesian estimation is feasible.Compared to direct sparseness inversion algorithm,precondition conjugate gradient sparseness inversion algorithm is stable,with higher accuracy and faster converging speed.