A method includes receiving, via a processor, input data based upon received seismic data, migrating, via the processor, the input data via a pre-stack depth migration technique to generate migrated input data, encoding, via the processor, the input data via an encoding function as a migration attribute to generate encoded input data having a migration function that is non-monotonic versus an attribute related to the input data, migrating, via the processor, the encoded input data via the pre-stack depth migration technique to generate migrated encoded input data, and generating an estimated common image gather based upon the migrated input data and the migrated encoded input data. The method also includes generating a seismic image utilizing the estimated common image gather, wherein the seismic image represents hydrocarbons in a subsurface region of the Earth or subsurface drilling hazards.

The present disclosure relates to a method of processing seismic signals comprising: receiving a set of seismic signals, applying a wavelet transformation to the set of signals and generating transformed signals across a plurality of scales. Then for each scale determining coherence information indicative of the transformed signals and generating a comparison matrix comparing the transformed signals, then outputting seismic attribute information based on combined coherence information.

A simultaneous sources seismic acquisition method is described that introduces notch diversity to improve separating the unknown contributions of one or more sources from a commonly acquired set of wavefield signals while still allowing for optimal reconstruction properties in certain diamond-shaped regions. In particular, notch diversity is obtained by heteroscale encoding.

Prestack images from the object store are hierarchically combined to generate a hierarchically stacked image. The hierarchically stacked image is generated by combining stacked images that includes a stacked image. The stacked image is generated by combining at least the prestack images. Based at least on the hierarchically stacked image, a quality measure of a prestack image is generated. Prior to deleting at least a subset of the prestack images from the object store and based at least on the quality measure, the prestack images are further combined to generate an enhanced stacked image. The stacked image is substituted using the enhanced stacked image. Subsequent to the substituting and prior to deleting at least the subset of the stacked images from the object store, the stacked images are combined to generate an enhanced hierarchically stacked image. The enhanced stacked image and the enhanced hierarchically stacked image are generated using failure recovery metadata. The enhanced hierarchically stacked image is presented.

Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for training a seismic data de-blending model. In one aspect, a method comprises: obtaining a plurality of de-blending training examples, wherein each de-blending training example defines: (i) one or more blended seismic traces, and (ii) for each blended seismic trace, a corresponding plurality of target unblended seismic traces; using the de-blending training examples to train a de-blending model having a plurality of de-blending model parameters, comprising, for each de-blending training example: processing the one or more blended seismic traces of the training example using the de-blending model to generate an output which defines, for each of the one or more blended seismic traces of the training example, a corresponding plurality of estimated unblended seismic traces; and adjusting values of the plurality of de-blending model parameters.

A method for deblending seismic signals includes entering as input to a computer recorded signals comprising seismic energy from a plurality of actuations of one or more seismic energy sources. A model of deblended seismic data and a blending matrix are initialized. A blending matrix inversion is performed using the initialized model. The inversion includes using a scaled objective function. The inversion is constrained by a thresholding operator. The thresholding operator is arranged to recover coefficients of the model of the deblended seismic data that are substantially nonzero, against a Gaussian white noise background. The thresholded model is projected into data space. Performing the blending matrix inversion is repeated if a data residual exceeds a selected threshold and the inversion is terminated if the data residual is below the selected threshold. At least one of storing and displaying an output of the blending matrix inversion is performed when the blending matrix inversion is terminated.

Methods and systems to separate seismic data associated with impulsive and non-impulsive sources are described. The impulsive and non-impulsive sources may be towed through a body of water by separate survey vessels. Receivers of one or more streamers towed through the body of water above a subterranean formation generate seismic data that represents a reflected wavefield produced by the subterranean formation in response to separate source wavefields generated by simultaneous activation of the impulsive source and the non-impulsive source. Methods and systems include separating the seismic data into impulsive source seismic data associated with the impulsive source and non-impulsive source seismic data associated with the non-impulsive.

The subject disclosure relates to the interpretation of borehole sonic data using machine learning. In one example of a method in accordance with aspects of the instant disclosure, borehole sonic data is received, and machine learning is used to interpret the borehole sonic data.

An apparatus, method, and non-transitory computer readable medium that can mitigate wireless channel impairments in seismic data transmission using deep neural networks is disclosed. The apparatus includes a receiving circuitry to receive seismic data and a processing circuitry. The processing circuitry is configured to apply a blind system identification process to the seismic data to estimate a channel impulse response of the seismic data, apply an optimum equalization process to obtain estimated seismic data based on the channel impulse response, process the estimated seismic data to generate processed seismic data, classify the processed seismic data into a first group of seismic data each of which has a signal-to-noise ratio (SNR) less than an SNR threshold and a second group of seismic data each of which has an SNR no less than the SNR threshold, and enhance the SNR of each of the first group of seismic signals.

Systems and methods are provided for a horizon patch extraction process and in particular, to receiving seismic trace data of a plurality of seismic events of a subterranean volume, selecting a first seismic trace based on the seismic trace data of the plurality of seismic events, the first seismic trace including a plurality of seismic onsets, determining a depth, an amplitude, and a first thickness of a first seismic onset of the first seismic trace, determining a second thickness between the first seismic onset and a second seismic onset, determining a third thickness between the first seismic onset and a third seismic onset, and generating a horizon patch based on the depth, the amplitude, and the first thickness of the first seismic onset, the second thickness between the first seismic onset and the second seismic onset, and the third thickness between the first seismic onset and the third seismic onset.