Seismic data enhancement
Various implementations described herein are directed to methods for processing seismic data including enhancing seismic data generated by a seismic sensor. The methods may include receiving seismic data that had been acquired using the seismic sensor. The seismic data may include one or more first seismic data elements related to a first seismic event. The methods may include generating summands of the seismic data as an inner product of multiple seismic data elements associated with the first seismic data elements and one or more second seismic data elements. The second seismic data elements may be different than the first seismic data elements. The methods may further include enhancing the seismic data by estimating outliers in the summands of the inner product and replacing the outliers with interpolated values.
1. A method for generating a seismic image, the method comprising:
receiving seismic data of a subsurface region wherein the seismic data are acquired by at least one seismic sensor;
performing an inversion approach via a conjugate gradient solver to separate two different portions of the seismic data, wherein the inversion approach comprises enhancing orthogonality of the two different portions of the seismic data for the conjugate gradient solver by analyzing components associated with an inner product of the two different portions of the seismic data and, based at least in part on the analyzing, replacing one or more of the components; and
based at least in part on at least one of the separate two different portions of the seismic data, generating a seismic image of subsurface reflectors of the subsurface region.
2. The method of claim 1 , wherein the seismic data comprises a series of samples of one or more seismic signals generated with at least one of the at least one seismic sensor, and wherein first seismic data elements comprise a first series of samples of a first seismic signal associated with a first seismic event.
3. The method of claim 1 , wherein the two different portions of the seismic data comprise first and second seismic data elements that comprise at least one of scalars, vectors, filters, and matrix components.
4. The method of claim 1 , wherein the two different portions of the seismic data comprise
one or more second seismic data elements; that are related to a second seismic event different than a first seismic event, and
a second series of samples of a second seismic signal associated with the second seismic event.
5. The method of claim 4 , wherein the first seismic event comprises a first shot fired by a seismic source, and wherein the second seismic event comprises a second shot fired by the seismic source after firing the first shot.
6. The method of claim 5 , wherein the second seismic data elements comprise extraneous signal noise derived from reflection of the first seismic event.
7. The method of claim 1 comprising generating summands of the seismic data by applying an inner product algorithm to the seismic data to generate the summands of the seismic data as the inner product of multiple seismic data elements associated with the two different portions of seismic data.
8. The method of claim 1 comprising generating summands of the seismic data by applying linear algebra to the seismic data to generate the summands of the seismic data as the inner product of multiple seismic data elements associated with the two different portions of seismic data.
9. The method of claim 1 comprising estimating outliers in summands of the inner product by identifying one or more largest values and one or more smallest values in the summands, wherein the one or more largest and smallest values are determined from either signed values or absolute values.
10. The method of claim 1 comprising removing outliers and replacing the removed outliers with interpolated values.
11. The method of claim 10 , wherein removing the outliers comprises zeroing one or more magnitudes of the largest values and one or more magnitudes of the smallest values in summands.
12. The method of claim 1 comprising replacing outliers with interpolated values by scaling the seismic data by replacing the outliers with a mean value between neighboring summands.
13. The method of claim 1 comprising replacing outliers with interpolated values by scaling the seismic data by replacing the outliers with an average value of remaining summands.
14. A device to generate a seismic image, the device comprising:
a processor; and
memory having instructions stored thereon that, when executed by the processor, cause the device to:
receive seismic data of a subsurface region wherein the seismic data are acquired by at least one seismic sensor;
perform an inversion approach via a conjugate gradient solver to separate two different portions of the seismic data, wherein the inversion approach enhances orthogonality of the two different portions of the seismic data for the conjugate gradient solver by analysis of components associated with an inner product of the two different portions of the seismic data and, based at least in part on the analysis, replaces one or more of the components; and
based at least in part on at least one of the separate two different portions of the seismic data, generate a seismic image of subsurface reflectors of the subsurface region.
15. The device of claim 14 , wherein the conjugate gradient solver performs
conjugate gradient iterations.
16. The device of claim 14 comprising instructions for replacing outliers with interpolated values by scaling the seismic data by replacing the outliers with a mean value between neighboring summands, replacing the outliers with an average value of remaining summands, or replacing the outliers with a Gaussian weighting of remaining summands.
17. The device of claim 14 , wherein the conjugate gradient solver implements a conjugate gradient (CG) iterative scheme.
18. The device of claim 14 , wherein the inversion approach is part of a migration application or a simultaneous source separation application.
19. A non-transitory computer-readable medium having stored thereon computer-executable instructions which, when executed by a computer, cause the computer to generate a seismic image, wherein the computer-executable instructions further cause the computer to:
receive seismic data of a subsurface region wherein the seismic data are acquired by at least one seismic sensor;
perform an inversion approach via a conjugate gradient solver to separate two different portions of the seismic data, wherein the inversion approach enhances orthogonality of the two different portions of the seismic data for the conjugate gradient solver by analysis of components associated with an inner product of the two different portions of the seismic data and, based at least in part on the analysis, replaces one or more of the components; and
based at least in part on at least one of the separate two different portions of the seismic data, generate a seismic image of subsurface reflectors of the subsurface region.