A machine-implemented method, at least one non-transitory computer-readable medium storing instructions, and a computing system are provided for attenuating noise. A computing system receives a seismic image and generates a first image using a first neural network configured to identify low-frequency ground roll in a seismic image, and a second image using a second neural network configured to identify reflections in the seismic image. A combined image is generated by combining the first image and the second image. The first neural network and the second neural network are adjusted to reduce a difference between the combined image and the seismic image using frequency constraint to guide separation of the seismic image into the first image and the second image.

A method for performing a formation-related operation based on corrected vertical seismic profile (VSP) data of an earth formation includes performing a VSP survey and applying a normal moveout (NMO) correction equation to the survey data that is a function of source offset to wellhead. The method also includes solving the NMO correction equation using a simulated annealing algorithm having an object function that is a coherence coefficient of semblance analysis of an NMO corrected reflection event within a time window to provide NMO corrected data. The method further includes performing the formation-related operation at at least one of a location, a depth and a depth interval based on the VSP NMO corrected data.

One method interpolates simulated seismic data of a coarse spatial sampling to a finer spatial sampling using a neural network. The neural network is previously trained using a set of simulated seismic data with the finer spatial sampling and a subset thereof with the coarse spatial sampling. The data is simulated using an image of the explored underground formation generated using real seismic data. The seismic dataset resulting from simulation and interpolation is used for denoising the seismic data acquired over the underground formation. Another method demigrates seismic data at a sparse density and then increases density by interpolating traces using a neural network.

The present application provides a vector denoising method and a vector denoising device for multicomponent seismic data, which relate to the field of seismic data processing technologies. The vector denoising method for multicomponent seismic data includes: decomposing multicomponent seismic gather data to obtain a plurality of small multicomponent seismic data; obtaining quaternary frequency domain seismic data by performing a quaternary Fourier transformation according to each of the plurality of small multicomponent seismic data; extracting frequency slices from the quaternary frequency domain seismic data in a quaternary frequency domain, and filtering the frequency slices by using a Cadzow filtering method to obtain filtered quaternary frequency domain seismic data; and performing an inverse quaternary Fourier transformation on the filtered quaternary frequency domain seismic data to obtain filtered seismic data of each component.

Processes and systems for generating images of a subterranean formation from recorded seismic data obtained in a marine survey are described. Processes and systems compute reverse-time receiver-motion-corrected upgoing and downgoing pressure wavefields at different locations of corresponding upgoing and downgoing observation levels based on the recorded seismic data. The reverse-time receiver-motion-corrected upgoing and downgoing pressure wavefields are time forwarded and extrapolated to obtain a corresponding receiver-motion-corrected upgoing and downgoing pressure wavefields at locations of a static observation level. An image of the subterranean formation is generated based at least in part on the receiver-motion-corrected upgoing pressure wavefield and the receiver-motion-corrected downgoing pressure wavefield.

A system for seismic imaging of a subterranean geological formation, the system includes a receiver configured to obtain seismic data comprising a data volume representing a post-stacked image. The system includes a filtering module configured to: apply frequency-wavenumber (F-K) filter to the data volume extract a negative-dip structure image and apply the frequency-wavenumber (F-K) filter to the data volume extract a positive-dip structure image. The system includes a diffraction rendering module configured to: multiply the positive-dip structure image with the negative-dip structure image and generate a diffraction-enhanced seismic image representing a geological formation of the data volume.

A method includes receiving modelled seismic data that is to be recognized by the at least one classification and/or segmentation processor. The modelled seismic data can be represented within a transform domain. The method includes generating an output via the at least one processor based on the received modelled seismic data. The method also includes comparing the output of the at least one processor with a desired output. The method also includes modifying the at least one processor so that the output of the processor corresponds to the desired output.

A method includes receiving modelled seismic data that is to be recognized by the at least one classification and/or segmentation processor. The modelled seismic data can be represented within a transform domain. The method includes generating an output via the at least one processor based on the received modelled seismic data. The method also includes comparing the output of the at least one processor with a desired output. The method also includes modifying the at least one processor so that the output of the processor corresponds to the desired output.

Seismic exploration methods and data processing apparatuses employ a deep neural network to remove seismic interference (SI) noise. Training data is generated by combining an SI model extracted using a conventional model from a subset of the seismic data, with SI free shots and simulated random noise. The trained DNN is used to process the entire seismic data thereby generating an image of subsurface formation for detecting presence and/or location of sought-after natural resources.

Recorded seismic data includes seismic traces having respective source orientation angles, where the source orientation angles represent deviations in seismic source orientation relative to an inline survey direction. A plurality of designature operators corresponding to respective designature orientation angles within a defined set of designature orientation angles may be generated. For a given member of the defined set of designature orientation angles, a corresponding designature operator may be applied to the recorded seismic data to generate designatured seismic data for the given designature orientation angle. For a given seismic trace having a given source orientation angle, the designatured seismic data may be interpolated to generate a designatured version of the given seismic trace. The results may be stored in a tangible, computer-readable medium.