基于非平稳滤波算子的最小二乘反射系数估计及宽带波阻抗成像

油气地震勘探的终极目标是定量估计地下介质参数进行含油气评价，经典全波形反演为代表的反演成像方法试图通过数据匹配来实现地下参数的定量估计，进而实现对储层的定量描述，是勘探地震理论研究的前沿阵地。然而，实际地震波场过于复杂使得经典全波形反演在理论和应用上都面临挑战。本次研究回避全波场到宽带地下参数的直接非线性投影，通过更稳健的最小二乘偏移估计高波数的垂直入射反射系数并将其表征成相对阻抗，再结合最优化的背景速度表征的背景阻抗实现绝对阻抗估计。其核心思想为通过分尺度、多步骤逐步线性化处理来进行不同波数带的的参数估计，并最终统一在宽带阻抗反演成像上。其核心技术为像域Hessian矩阵的估计及其保真反射系数估计、反演框架下利用总变差正则化联合背景约束实现相对阻抗和背景阻抗的有机融合。项目旨在反演成像框架下探索一条区别于全波形反演的更为稳健的宽波数带波阻抗估计路线，并应用于实际资料处理。

The ultimate goal of oil and gas seismic exploration is to estimate the subsurface parameters quantitatively. The inversion imaging method, such as full waveform inversion, attempts to achieve quantitative estimation of subsurface parameters through data fitting, and then realize the reservoir description. It is the hotspot of theoretical research on seismic exploration. However, the field seismic wave field is too complicated, making the classical full waveform inversion encounter challenges in both theory and application aspects. This study avoids the direct nonlinear projection from the full wave field to the broadband subsurface parameters, and obtain the high-wavenumber reflection coefficient by a more robust least squares migration. Since the reflection coefficient has the some component of a relative impedance, the absolute impedance can be achieved straightforward by combining with the background impedance produced by background velocity. The key idea of this proposal can be expressed by implementing broadband parameter estimation by stepwise linearization strategy in different scales. The core technologies are proceeding a fidelity reflection coefficient estimation by introducing the approximate the Hessian matrix in imaging domain, estimenting the absolute impedance by taking cosider of the total variation regularization and background constraints in the inversion framework. The proposal aims to explore a more robust broadband impedance estimation route that is different from full waveform inversion under the imaging framework, and apply it to field data processing.