Abstract: A method for hydrocarbon exploration based on imaging tunnel valleys is disclosed. The method includes obtaining a 3D seismic volume data corresponding to a target formation having at least one tunnel valley, interpreting a key horizon at or above the target formation as preparation for paleo-depositional environment restoration, flattening and decimating the 3D seismic volume data using the key horizon for paleo-depositional environment restoration to obtain a conditioned 3D seismic volume data, analyzing the conditioned 3D seismic volume data for frequency content and decomposing the conditioned 3D seismic volume data into at least three attributes, blending the at least three attributes to form a single seismic volume data to illuminate key features, and displaying, on a map, a distribution of the tunnel valleys in the 3D seismic volume data of the target formation.

Abstract: Methods and systems, including computer programs encoded on a computer storage medium can be used to preserve edges while performing Gaussian grid smoothing of noise components in subsurface grids to generate geological maps. A subsurface grid is generated from data indicating properties of subsurface formations. A weighting grid is generated by: i) receiving seismic data representing the subsurface formations; ii) generating seismic attributes associated with discontinuities in the subsurface formations; and iii) assigning a particular weight value to weighting grid points that the seismic attributes associated with discontinuities in the subsurface formations indicate the presence of a discontinuity. The subsurface grid is processed by iteratively computing local averages of grid points in the subsurface grid using a compact Gaussian filter weighted by values in the weighting grid. A geological map of subsurface formations is generated based on the filtered subsurface grid.

Abstract: A method for hydrocarbon exploration based on imaging tunnel valleys is disclosed. The method includes obtaining a 3D seismic volume data corresponding to a target formation having at least one tunnel valley, interpreting a key horizon at or above the target formation as preparation for paleo-depositional environment restoration, flattening and decimating the 3D seismic volume data using the key horizon for paleo-depositional environment restoration to obtain a conditioned 3D seismic volume data, analyzing the conditioned 3D seismic volume data for frequency content and decomposing the conditioned 3D seismic volume data into at least three attributes, blending the at least three attributes to form a single seismic volume data to illuminate key features, and displaying, on a map, a distribution of the tunnel valleys in the 3D seismic volume data of the target formation.

Abstract: Methods and systems, including computer programs encoded on a computer storage medium can be used to preserve edges while performing Gaussian grid smoothing of noise components in subsurface grids to generate geological maps. A subsurface grid is generated from data indicating properties of subsurface formations. A weighting grid is generated by: i) receiving seismic data representing the subsurface formations; ii) generating seismic attributes associated with discontinuities in the subsurface formations; and iii) assigning a particular weight value to weighting grid points that the seismic attributes associated with discontinuities in the subsurface formations indicate the presence of a discontinuity. The subsurface grid is processed by iteratively computing local averages of grid points in the subsurface grid using a compact Gaussian filter weighted by values in the weighting grid. A geological map of subsurface formations is generated based on the filtered subsurface grid.

Abstract: Methods for machine learning-based differencing for hydrocarbon well logs are disclosed. A computer system generates an input layer of a machine learning module. The input layer includes a graphical representation of datasets obtained from one or more hydrocarbon wells. The computer system generates an encoding layer of the machine learning module from the graphical representation. The encoding layer includes a two-dimensional array based on the multiple datasets and a merged dataset of the one or more hydrocarbon wells. The computer system determines that a difference between a first hydrocarbon well log of the multiple datasets and a second hydrocarbon well log of the merged dataset warrants an action. The determining is performed using a convolutional layer. Responsive to determining that the difference warrants an action, the computer system performs the action on the first hydrocarbon well log and the second hydrocarbon well log to modify the merged dataset.