Our method of resolving local dips into time shifts has effectively flattened seismic in our test cases.
Our use of the 3D Fourier transform may not be necessary. We maybe able to integrate the dips in 2D. This would make this method capable of handling large data sets easily. As mentioned earlier, we can smooth both the numerator and denominator of the dip calculation in equation (1) along all three axis. This could possibly eliminate the need for integrating in the t direction by properly smoothing the dip calculation.
The ability of this method to work with data that has pinch outs and faults still needs to be looked at. A local dip estimator will estimate incorrect dips at faults. Compounding the problem, our dip integration method will try to honor those incorrect dips. Once in the Fourier domain, it will be very hard to correct this problem.
Overall, the results of this method are very encouraging. The ability to flatten data could be a powerful tool in automating interpretation in general. There could be many processing applications as well, such as flattening gathers.