A method for training a backpropagation-enabled regression process is used for predicting values of an attribute of subsurface data. A multi-dimensional seismic data set with an input dimension of at least two is inputted into a backpropagation-enabled process. A predicted value of the attribute has a prediction dimension of at least 1 and is at least 1 dimension less than the input dimension.

A method includes receiving a training selection of a first set of faults located in a first subset of a seismic dataset for a subsurface geologic formation, detecting a second set of faults in the seismic dataset based on fault interpretation operations using a first set of interpretation parameters, and determining a difference between the first set of faults and the second set of faults. The method also includes generating a second set of interpretation parameters for the fault interpretation operations based on the difference between the first set of faults and the second set of faults, and determining a feature of the subsurface geologic formation based on fault interpretation operations using the second set of interpretation parameters.

A methodology for discontinuous smooth interpolation in order to generate a curve of a discontinuous volume due to one or more faults in a subsurface is disclosed. Faults in a subsurface result in discontinuities in the subsurface. Hydrocarbon management may seek to determine various surfaces in the subsurface, including across the faults in the subsurface. To generate the various surfaces, a continuous formulation of the interpolation method is followed in which discontinuous smooth interpolation is viewed as a variational optimization problem (such as an energy optimization problem) for the surface curvature function. In this way, the methodology does not require that the input data be located at grid points and discretized with a structured regular grid. Rather, because a continuous function is used, an unstructured grid may also be used to discretize the resulting equation.

Well information may define subsurface configuration of different wells. Marker information defining marker positions within the wells may be obtained. A dissimilarity matrix for the wells may generated, with the element values of the dissimilarity matrix determined based on comparison of corresponding subsurface configuration of the wells. A gated dissimilarity matrix may be generated from the dissimilarity matrix based on the marker positions within the wells. The elements values of the gated dissimilarity matrix corresponding to one set of marker positions and not corresponding to the other set of marker positions may be changed. Correlation between the wells may be determined based on the gated dissimilarity matrix such that correlation exists between a marker position in one well and a marker position in another well.

A method can include receiving seismic image data of a geologic region and interpretation information of the seismic image data for a geologic feature in the geologic region, where the seismic image data include a geologic feature class imbalance; shifting the geologic feature class imbalance toward a class of the geologic feature by increasing the spatial presence of the geologic feature in the seismic image data to generate training data; and training a neural network using the training data to generate a trained neural network.

A computer executable algorithm adapted to propagate a boundary surface of a seed that is placed within a region of interest of a visual representation of a 3D seismic data so as to follow a natural contour of said region of interest, wherein said algorithm is executable to: (i) generate at least one attribute volume comprising at least on attribute derivable from said 3D seismic data set; (ii) generate at least one characteristic parameter for a plurality of candidate events of said 3D seismic data within a predefined gate region located forward of said propagating boundary surface; (iii) generate and assign a probability characteristic for said plurality of candidate events based on said at least one attribute volume and said at least one characteristic parameter; and propagate said boundary surface towards and incorporating any one of said plurality of candidate events that fulfils an acceptance criteria of said probability characteristic so as to generate a surface along the natural contour of said region of interest.

The invention discloses a dolomite reservoir prediction method and system based on well and seismic combination, and storage medium. The method steps include: obtaining the dolomite index characteristic curve through well log sensitivity analysis, and distinguishing the dolomite and limestone according to the difference in their response range; after the artificial intelligence deep learning is performed on the dolomite index characteristic curve of the drilling area, the dolomite index characteristic curve of the virtual drilling area is obtained; according to the dolomite index characteristic curve of the drilling area and the virtual drilling area, the post-stack seismic data is used for inversion to obtain the distribution and development status of the dolomite reservoir in the test area. The invention effectively distinguishes the dolomite and limestone through the dolomite index characteristic curve, and accurately predicts the distribution and development status of the dolomite reservoir in the test area with less wells.

Methods, computing systems, and computer-readable media for interpreting seismic data, of which the method includes receiving seismic data representing a subterranean volume, and determining a feature-likelihood attribute of at least a portion of a section of the seismic data. The feature-likelihood attribute comprises a value for elements of the section, the value being based on a likelihood that the element represents part of a subterranean feature. The method also includes identifying contours of the subterranean feature based in part on the feature-likelihood attribute of the section, and determining a polygonal line that approximates the subterranean feature.

A method is described for seismic imaging including image enhancement using a trained neural network. The neural network may receive training pairs of low signal-to-noise ratio 3D seismic images and high signal-to-noise ratio 3D seismic images; train a neural network on the training pairs wherein the training uses atrous convolution; receive a seismic image representative of a subsurface volume of interest; apply the neural network to the seismic image to generate a second seismic image; and display the second seismic image on a graphical user interface. The method is executed by a computer system.

The present disclosure discloses a method for identifying a boundary of a sedimentary facies, a computer device and a computer readable storage medium. The method comprises: acquiring a preliminary marked result of the sedimentary facies in a seismic attribute map; acquiring a color-based K-means classification result of the seismic attribute map by using a maximal between-cluster variance and a K-means clustering; acquiring a super-pixel classification result of the seismic attribute map according to a SLIC super-pixel segmentation; and performing a region growing fusion on the super-pixel classification result by taking the preliminary marked result and the K-means classification result as constraints, to determine an identification result of the boundary of the sedimentary facies in the seismic attribute map.