基于低频软约束的叠前AVA稀疏层反演

AVA sparse layer inversion with the soft-low frequency constraint

叠前AVA楔形字典利用基追踪分解算法实现叠前AVA稀疏层反演,较常规叠前反演具有较高的垂向分辨率,但由于该算法没有考虑弹性参数之间的相关性引起的不适定性,造成密度反演结果非常不稳定。为此,基于Zhang的常规叠前AVA稀疏层反演,结合Downton的参数协方差矩阵特征分解去相关思路,重新推导了叠前AVA楔形字典和叠前AVA稀疏层反演的目标函数;在此基础上,通过引入低频软约束项补偿反演的低频信息,可以获取全频带的反演结果,进一步降低了三参数反演的不适定性,提高了密度项的反演精度。模型数据和实际数据试算结果表明,基于低频软约束的叠前AVA稀疏层反演具有较高的稳定性和垂向分辨率。

Using the AVA wedge dictionary for the basic pursuit decomposition,the AVA sparse layer inversion owns high resolution.However conventional AVA sparse layer inversions do not take the correlation among the elastic parameters into account,and result in the ill-posedness.Thus the density inversion result is very unstable.So we propose an AVA sparse layer inversion with the soft-low frequency constraint.First the AVA wedge dictionary and the objective function of AVA sparse layer inversion are rearranged by combined the decorrelation based on the eigendecomposition of the covariance matrix proposed by Downton.On this basis,by introducing the soft low-frequency constraint,the low frequency component of the inversion can be compensated and the full band inversion result can be acquired,which results in reducing the ill-posedness and improving the accuracy of the inverted density.Model test and real data application prove that the AVA sparse layer inversion with the soft low-frequency constraint owns the high stability and the high resolution.