时频域振幅相位联合的最小二乘逆时偏移

最小二乘逆时偏移方法具有复杂地质构造成像精度高、成像振幅准确等优点.但是,当地下存在强散射介质时,最小二乘逆时偏移方法很难透过上覆强散射地质体获得深部构造的高精度成像结果.本文为了提高深部精细构造的成像质量,提出时频域振幅相位联合的最小二乘逆时偏移方法.该方法主要通过构建时频域振幅相位联合目标函数,减弱振幅信息对成像结果的影响,提高深部弱散射地震信号的可成像精度.首先,对地震信号进行时频变换,构建时频域最小二乘偏移目标函数;其次,在目标函数中引入振幅权重因子,调节时频域振幅相位权重;最后,推导时频域振幅相位联合目标函数对模型参数的梯度,并利用L-BFGS局部优化算法对成像结果进行迭代.Marmousi模型和盐丘模型测试结果表明,本文方法能够很好地利用弱散射地震信号的时频域振幅相位信息,实现透过上覆强散射地质体进行深部高精度成像的目标.

Joint least square reverse time migration of phase and amplitude in the time-frequency domain

Least Square Reverse Time Migration (LSRTM) has the advantages of high-resolution imaging and accurate amplitude in complex geological structures. However, if subsurface exist strong-scattering bodies, this method is difficult to penetrate these overlying media to obtain high-resolution imaging results. To improve the imaging quality of deep structures, a joint LSRTM of Phase and Amplitude Method (PA-LSRTM) in the time-frequency domain is proposed in this paper. It establishes a time-frequency phase and amplitude misfit to weaken the effect of amplitude components on the imaging results, thus the imaging resolution of deep weak scattering seismic signal can be improved. Firstly, the shot recorded seismic data are transformed into that in the time-frequency domain to establish a PA-LSRTM misfit. Then, an amplitude factor is introduced in the PA-LSRTM misfit to adjust the weight of time-frequency phase and amplitude. Finally, it takes derivative with respect to the model parameter, and uses the L-BFGS optimization algorithm to update the imaging result. The numerical tests on Marmousi and salt models demonstrate that the PA-LSRTM can take good use of the time-frequency domain phase and amplitude information to penetrate overlying strong scattering bodies and improve the imaging results of sub-bottom and sub-salt structures.