《Non-Linear Least-Squares Reverse-Time Migration》.pdf

/ Non-Linear Least-Squares Reverse-Time Migration g r Gang Yao* and Helmut Jakubowicz, Imperial College London o . g e s. Summary source location, and is angular frequency. The recorded y r data, , are then back-propagated into the subsurface using a r We present a new implementation of least-squares reverse- the adjoint of the modeling operator, , giving b i l time migration which accounts for the non-linear / / : relationship between the data and reflectivity. As with p ( | ) ( ) (,)= ; , . (2) t t other least-squares migration methods, the key feature of h t this approach is that it uses an approximation for the a The imaging step generates an estimate of the reflectivity, e inverse of the modeling operator instead of its adjoint. (), by relating and to each other at all points in the s U However, in applying reverse-time migration, the two-way f subsurface. This is usually done by correlating the o wave equation is used, which in turn