一种新的针对反射波的逆时偏移真振幅成像条件

真振幅成像是一种代表性的定量估计模型参数扰动高波数部分的地震波成像方法.经典的真振幅成像方法在高频近似和理想照明假设条件下求取显式对角Hessian逆矩阵作为偏移振幅加权算子，用以校正波传播过程中的几何扩散效应，得到模型参数扰动的带限估计.真振幅保真成像方法在利用逆时偏移（RTM）框架实现时会产生低波数噪声，影响对高波数参数估计的精度.本文给出了一种新的基于RTM框架的真振幅保真成像条件，该成像条件针对反射波数据，在高频近似下散射模式对应正问题及Bayes反问题框架下导出.与传统基于高频渐进反演的波动方程成像方法类似，利用本文提出RTM成像条件能够保证计算结果与高频近似下反演结果的一致性.同时，利用本文提出RTM真振幅成像条件能够在成像过程中自动保真的消除传统真振幅RTM算法中存在低波数噪声，模型数值实验结果验证了本文方法的正确性和有效性.

A new true-amplitude imaging condition of reverse time migration for reflected seismic data

True-amplitude imaging is one of the typical seismic imaging method for quantitatively imaging high-wavenumber model perturbation. Under high-frequency and perfect illumination assumptions, the classical true-amplitude imaging method adopt an explicit inverse diagonal Hessian for removing geometric spreading during wave propagation, aiming at obtaining band-limited model perturbation. When implementing true-amplitude imaging through a RTM (Reverse Time Migration) process, the unwanted low-wavenumber artifacts are produced, which will contaminate the imaging accuracy for true seismic reflectors. In this paper, we propose new inversion imaging condition based on RTM process. The present RTM imaging condition aims at imaging reflected seismic data and derived with linearized surface scattering model, together with Bayes inverse problem framework. As the conventional asymptotic inversion based wave-equation imaging methods, the output of new RTM imaging condition is consistent with the asymptotic inversion in high frequency sense. In the meantime, the proposed imaging condition automatically suppress the RTM artifacts in a way that the phase and amplitude of the target reflector image is maintained. Numerical examples on synthetic data verify effectiveness of the present RTM imaging condition.