Techniques for processing of seismic data. A seismic data set is received, wherein the seismic data set comprises a first data subset associated with a first seismic source and a second data subset associated with a second seismic source. An input is received indicating that a distance between the first seismic source and the second seismic source is greater than or equal to a threshold value. The second data set is filtered from the seismic data set to remove the second data subset from seismic data set to generate a filtered seismic data set in response to receiving the input and a coherence volume is generated based on the filtered seismic data set.

Methods are presented for determining the location of underground features (e.g., CO.sub.2). One method includes capturing, by sensors distributed throughout a region, seismic traces associated with seismic signals generated by a seismic source. For multiple sensors, active noise is identified or passive noise is measured within each seismic trace and values for attributes associated with the active or passive noise are determined. Further, an unsupervised machine-learning model, based on the values of the attributes, is utilized to determine noise characteristics for multiple sensors. The sensors are grouped in clusters based on the noise characteristics for each sensor. For multiple clusters, a noise filter is created based on the noise characteristics of the sensors in the cluster, and the noise filter of the cluster is applied, for multiple sensors, to the seismic traces of the sensor. Additionally, the filtered seismic traces are analyzed to determine a location of CO.sub.2 underground.

The present disclosure describes methods and systems, including computer-implemented methods, computer program products, and computer systems, for suppressing noises in seismic data. One computer-implemented method includes receiving, at a data processing apparatus, a set of seismic data associated with a subsurface region; flattening, by the data processing apparatus, the set of seismic data according to an identified seismic event; dividing, by the data processing apparatus, the set of seismic data into a plurality of spatial windows; randomizing, by the data processing apparatus, the set of seismic data according to a random sequential order; filtering, by the data processing apparatus, the randomized seismic data; and reorganizing, by the data processing apparatus, the filtered seismic data according to a pre-randomization order.

The present method predicts and separates dispersive surface waves from seismic data using dispersion estimation and is completely data-driven and computer automated and no human intervention is needed. The method is capable of predicting and suppressing surface waves from recorded seismic data without damaging the reflections. Nonlinear signal comparison (NLSC) is used to obtain a high resolution and accurate dispersion. Based on the dispersion, surface waves are predicted from the field recorded seismic data. The predicted surface waves are then subtracted from the original data.

Systems, methods, and computer-readable media for the attenuation of interface waves using polarization filtering applied to recorded single component seismic data are disclosed. A second component for polarization filtering is created by determining interface waves from the recorded data single component seismic data. The second component seismic data may be generated using an interface waves propagation model (in frequency or time-frequency domain) or by differential normal move-out (NMO) interpolation. Polarization filtering may be applied to multicomponent seismic data formed from the recorded single component seismic data and the generated second component seismic data to attenuate interface noise.

A system and method for attenuating surface waves in common shot gathers of seismic data recorded by a set of geophones by: iteratively executing a genetic algorithm over a plurality of generations to generate an optimal one-dimensional (1D) Earth model based on the common shot gather data by, successively refining a pool of candidate Earth models to better fit the common shot gather data, until optimal Earth models in sequential generations converge; generating synthetic surface wave data based on the optimal Earth model and canceling the synthetic surface wave data from the common shot gather data to generate new common shot gather data that reduces the noise due to surface waves; and iteratively executing the genetic algorithm over each new common shot gather data until optimal Earth models generated in sequential iterations of the genetic algorithm converge.

Systems, methods, and computer-readable media for attenuating guided waves in seismic data using polarization filtering are provided. A raw hydrophone component and raw geophone component of multicomponent seismic data may be scaled using a constant scalar to enhance the ellipticity ratio of guided waves. Polarization filtering based on the ellipticity ratio may be applied within a velocity constraint to the scaled hydrophone and vertical geophone components to attenuate the guided waves. Additionally or alternatively, polarization filtering based on the tilt angle may be applied within a velocity constraint to the raw hydrophone and vertical geophone components to attenuate the guided waves. Polarization filtering may be applied to a raw hydrophone component and raw vertical geophone component of seismic data to attenuate Scholte waves before attenuation of the guided waves.

Techniques for processing of seismic data. A seismic data set is received, wherein the seismic data set comprises a first data subset associated with a first seismic source and a second data subset associated with a second seismic source. An input is received indicating that a distance between the first seismic source and the second seismic source is greater than or equal to a threshold value. The second data set is filtered from the seismic data set to remove the second data subset from seismic data set to generate a filtered seismic data set in response to receiving the input and a coherence volume is generated based on the filtered seismic data set.

A method for processing sonic data acquired with a downhole sonic tool is provided. The method comprises detecting coherent noise based on a plurality of waveforms obtained from one or more receivers issued by one or more transmitters. The plurality of waveforms correspond to propagating acoustic waves in a formation. In addition, the method comprises building a slowness filter for removing the coherent noise, and applying the slowness filter to the plurality of waveforms.

A method for removing a guided wave noise in a time-domain may include recording one or more acoustic signals with one or more receivers at a first location, wherein the one or more acoustic signals are raw data. The method may further include determining a slowness range, estimating a downward guided wave noise by stacking the one or more acoustic signals based at least in part on a positive slowness, estimating an upward guided wave noise by stacking the one or more acoustic signals based at least in part on a negative slowness, and identifying a dominant direction of propagation. The method may further include identifying a slowness from a highest stacked amplitude for the dominant direction of propagation, estimating a downward guided wave noise with the slowness, estimating an upward guided wave noise with the slowness, and subtracting the downward guided wave noise and the upward guided wave noise.