Device and method for model-based deblending

1. A method for removing cross-talk noise from seismic data and generating an image of a surveyed subsurface, the method comprising:

receiving input seismic data D generated by firing one or more seismic sources so that detected source energy is overlapping, and the input seismic data D is recorded with seismic sensors over the surveyed subsurface;

generating a cross-talk noise model N by removing at least one original gather and replacing the at least one original gather with a reconstructed gather that is generated based on neighboring gathers of the at least one original gather;

subtracting the cross-talk noise model N from the input seismic data D to attenuate coherent cross-talk noise to obtain processed seismic data DP;

deblending the processed seismic data DP with a deblending algorithm to attenuate a residual noise to obtain deblended seismic data D d ; and

generating the image of the subsurface based on the deblended seismic data D d .

2. The method of claim 1 , wherein the reconstructed gather is generated based on an interpolation of the neighboring gathers.

3. The method of claim 1 , wherein the neighboring gathers are original recorded seismic data.

4. The method of claim 1 , wherein the neighboring gathers are reconstructed gathers.

5. The method of claim 1 , wherein the neighboring gathers are a mixture of original recorded seismic data and reconstructed gathers.

6. The method of claim 1 , wherein the generating the cross-talk noise model N comprises:

removing plural original gathers besides the at least one original gather from the input seismic data D; and

generating plural reconstructed gathers to replace the plural original gathers, each of the plural reconstructed gathers being based on respective neighboring gathers of one of the plural original gathers that the each of the plural reconstructed gathers replaces.

7. The method of claim 6 , further comprising:

muting cross-talk noise parts of the reconstructed gathers to obtain a part with a highest signal-to-noise ratio for each one of the reconstructed gathers; and

applying time corrections to the part with the highest signal-to-noise ratio to obtain the cross-talk noise model, wherein the time corrections are related to shot firing times.

8. The method of claim 7 , wherein the time corrections are applied prior to the subtracting, on either the input seismic data or the cross-talk noise model or both.

9. The method of claim 8 , wherein the time corrections are related to either a primary shot or a blended shot.

10. The method of claim 1 , wherein the reconstructed gather is obtained using a simple interpolation, a sparse interpolation or an inversion method in a shot, a channel, or a common mid-point or by replacing a shot with an adjacent shot data.

11. The method of claim 1 , wherein the cross-talk noise model is subtracted from the input seismic data D in a shot domain, a channel domain or a common mid-point domain.

12. A computing device for removing cross-talk noise from seismic data and generating an image of a surveyed subsurface, the computing device comprising:

an interface configured to receive input seismic data D generated by firing one or more seismic sources so that detected source energy is overlapping, and the input seismic data D is recorded with seismic sensors over the surveyed subsurface; and

a processor connected to the interface and configured to:

generate a cross-talk noise model N by removing at least one original gather and replace the at least one original gather with a reconstructed gather that is generated based on neighboring gathers of the at least one original gather;

subtract the cross-talk noise model N from the input seismic data D to attenuate coherent cross-talk noise to obtain processed seismic data D P ;

deblend the processed seismic data DP with a deblending algorithm to attenuate a residual noise to obtain deblended seismic data D d ; and

generate the image of the subsurface based on the deblended seismic data D d .

13. The computing device of claim 12 , wherein the reconstructed gather is generated based on an interpolation of the neighboring gathers.

14. The computing device of claim 12 , wherein the neighboring gathers are original recorded seismic data.

15. The computing device of claim 12 , wherein the neighboring gathers are reconstructed gathers.

16. The computing device of claim 12 , wherein the neighboring gathers are a mixture of original recorded seismic data and reconstructed gathers.

17. The computing device of claim 12 , wherein the processor is further configured to:

remove plural original gathers besides the at least one original gather from the input seismic data D; and

generate plural reconstructed gathers to replace the plural original gathers, each of the plural reconstructed gathers being based on respective neighboring gathers of one of the plural original gathers that the each of the plural reconstructed gathers replaces.

18. A non-transitory computer-readable medium storing instructions, which when executed by a processor, implement a method for removing cross-talk noise from seismic data and generating an image of a surveyed subsurface, the instructions causing:

receiving input seismic data D generated by firing one or more seismic sources so that detected source energy is overlapping, and the input seismic data D is recorded with seismic sensors over the surveyed subsurface;

generating a cross-talk noise model N by removing at least one original gather and replacing the at least one original gather with a reconstructed gather that is generated based on neighboring gathers of the at least one original gather;

subtracting the cross-talk noise model N from the input seismic data D to attenuate coherent cross-talk noise to obtain processed seismic data D P ;

deblending the processed seismic data D P with a deblending algorithm to attenuate a residual noise to obtain deblended seismic data D d ; and

generating the image of the subsurface based on the deblended seismic data D d .