Next: [4] Analysis of some
Up: Aspects of Linearized migration/inversion
Previous: [2] The matrix expression
The iterative formula of the leastsquares migration/inversion is:
 
(74) 
where H^{k+1} is the inverse of the Hessian. The firstorder derivative of the cost function with respect to the medium parameters is
 
(75) 
If the residual wavefield is defined as
 
(76) 
and equation (75) is rewritten as
 
(77) 
then the firstorder derivative means that the residual wavefield is backpropagated. It is further equivalent to the classical prestack migration if the parameter disturbance is set to zero at the first iteration. The residual wavefield belongs to the data space , and pertains to the image space .Calculating the firstorder derivative requires onetime modeling , which can be implemented by a prestack demigration, and onetime classical prestack migration of :
 

 (78) 
In the first iterative step, , equation (78) is rewritten as
 

 
 (79) 
The Hessian is the secondorder derivative of the cost function with respect to the medium parameters. It is of the following form:
 
(80) 
In the leastsquares migration/inversion, the Hessian is a deconvolution operator. It is used for deblurring the image of the classical prestack migration. Physically, the Hessian is an indicator of the illumination. The energy of the wave propagating through a certain medium is expressed as follows:
 
(81) 
For a given layer and from the modeling equation (58), equation (81) can be rewritten as
 

 (82) 
Clearly, for a horizontal reflector with an even reflectivity and only the zerooffset reflectivity considered, A^{H}A determines the energy of the wave which propagates to the layer. Equation (82) can be rewritten as follows:
 

 
 
 
 (83) 
The row in the matrix multiplied by the column of the matrix and the row in the matrix multiplied by the column of the matrix are the crosscorrelation between the conjugate of the Green's function and the Green's function at different receiver or shot positions respectively. The autocorrelation has a peak value, and the crosscorrelation decreases rapidly as the distance increases between the receiver and shot positions. The autocorrelation values are on the diagonal. Therefore, the Hessian is a bandwidthlimited diagonal matrix. Its inverse is also a kind of bandwidthlimited diagonal matrix. In the extreme case, where only the elements on the diagonal of the Hessian are left, with nondiagonal set to zero, the elements on the diagonal of the inverse of the Hessian are the reciprocals of the elements on the diagonal of the Hessian. Therefore, the inverse of the Hessian plays the role of decreasing strong illumination and enhancing poor illumination. The Hessian itself reflects the illumination of each imaging point.
The matrix expression of migration/inversion can be summarized as follows:
 

 
 
 (84) 
Substituting the residual imaging matrix into equation (84), it can be rewritten as follows:
 

 
 
 (85) 
With the quasiNewton condition , and , , the DFP algorithm for calculating the inverse of the Hessian matrix is
 
(86) 
where
 
(87) 
and
 

 
 (88) 
where r_{ij}^{residual} is the image with the residual wavefield.