next up previous print clean
Next: Explicit vs. implicit Hessian Up: Expanding Hessian dimensionality Previous: Subsurface-offset Hessian

Angle-domain Hessian

Sava and Fomel (2003) define an image space transformation from subsurface offset to reflection and azimuth angle as:
   \begin{eqnarray}
{\bf m}({\bf x},{\Theta}) &=& {\bf T'} (\Theta,{\bf h}){\bf m}({\bf x},{\bf h}), 

\end{eqnarray} (12)
where ${\Theta}=(\theta,\alpha)$ are the reflection and the azimuth angles, and ${\bf T'} (\Theta,{\bf h})$ is the adjoint of the angle-to-offset transformation operator (slant stack).

Substituting the prestack migration image (subsurface offset domain) in equation 7 into equation 12 we obtain the expression for the prestack migration image in the angle-domain that follows:
   \begin{eqnarray}
{\bf m}({\bf x},{\Theta})&=& {\bf T'} (\Theta,{\bf h}) {\bf L}'...
 ...{'}\sum_{{\bf x}}{\bf d}({\bf x}_s,{\bf x}_r;\omega). \nonumber

\end{eqnarray} (13)

The synthetic data can be modeled (as the adjoint of equation 14) by the chain of linear operator ${\bf L}$ and the angle-to-offset transformation operator acting on the model space,
   \begin{eqnarray}
{\bf d}({\bf x}_s,{\bf x}_r;\omega)&=&{\bf L}{\bf T} (\Theta,{\...
 ...m_{\omega}
 {\bf T} (\Theta,{\bf h}){\bf m}({\bf x},{\Theta}),

\end{eqnarray}
(14)
The quadratic cost function is
\begin{eqnarray}
S({\bf m}) &=& \frac{1}{2} \sum_{\omega}\sum_{{\bf x}_s}\sum_{{...
 ...\left[ {\bf d}({\bf x}_s,{\bf x}_r;\omega)-{\bf d}_{obs} \right],
\end{eqnarray}
(15)
while its first derivative with respect to the model parameters ${\bf m}({\bf x},\Theta)$ is
\begin{eqnarray}
\frac{\partial{S({\bf m})}}{\partial{{\bf m}({\bf x},\Theta)}}=...
 ...) {\bf G}({\bf x+h},{\bf x}_s;\omega){\bf T} (\Theta,{\bf h})
\}
\end{eqnarray}
(16)
and its second derivative with respect to the model parameters ${\bf m}({\bf x},\Theta)$ and ${\bf m}({\bf x'},\Theta')$ is the angle-domain Hessian
   \begin{eqnarray}
{\bf H}({\bf x,\Theta};{\bf x',\Theta'})&=&
\frac{\partial^2{S...
 ...bf h}){\bf H}({\bf x,h};{\bf x',h'}) {\bf T}(\Theta',{\bf h'}).

\end{eqnarray}
(17)


next up previous print clean
Next: Explicit vs. implicit Hessian Up: Expanding Hessian dimensionality Previous: Subsurface-offset Hessian
Stanford Exploration Project
10/31/2005