Image-Domain Least-Squares Reverse-Time Migration Using Point Spread Functions
Jing Bai, O. Yilmaz · 2021
Summary Reverse-time migration (RTM) produces image of subsurface structures. However, migration image can be distorted due to spatial aliasing, limit record aperture, illumination effects, and so on. To partially correct for the distortion effects, we present an image-domain least-squares reverse-time migration (LSRTM) by approximating the Hessian through point spread functions (PSFs) with a regularization of the difference between migration image and a reflectivity model and a total variation regularization of the reflectivity model. The optimization problem is iteratively solved by a nonlinear conjugate gradient method for an optimal reflectivity model. In order to improve image quality and subsalt imaging, a weighting function is necessarily applied to migration image for complex structures with salt or dirty salt bodies. Tests demonstrate that the proposed LSRTM is helpful to remove migration artifacts, improve image resolution and subsalt imaging in optimal reflectivity models.