|
<< Click to Display Table of Contents >> Navigation: MultiFocusing > Engine - 3D CO-MF |
Engine - 3D CO-MF is the core computation module for 3D Common-Offset MultiFocusing (CO-MF) parameter search. For each bin in the survey, the engine searches for the optimal set of MultiFocusing parameters that best focuses pre-stack seismic energy across common-offset gathers, accounting for wavefield curvature and acquisition geometry in three dimensions.
The module operates on binned pre-stack seismic data and produces a database file (KDB) containing the optimized MF parameters and corresponding time-corrected events for each computed bin. This database is then used by the Data Enhance modules to regularize and interpolate the seismic data. Because the output is already registered to the bin grid, no additional binning should be applied after data enhancement — doing so may introduce artifacts.
Computation can be distributed across CPU threads or GPU devices on a single workstation, or dispatched to a remote cluster for large 3D surveys. The engine supports incremental recalculation: previously completed bins are skipped automatically unless explicitly marked for reprocessing, making it efficient to rerun only updated areas of the survey.
Connect to the SEG-Y file handle containing the pre-stack seismic data. This is the primary amplitude data source from which the engine reads traces for MF parameter search. Visible only when Use trace vector on disk is disabled.
Connect to the sorted trace headers (index) for the input seismic data. These headers provide source-receiver geometry information (inline, crossline, offset, XY coordinates) used by the geometry extractor to locate traces within the CO aperture of each bin. Visible only when Use trace vector on disk is disabled.
Connect to the bin grid item that describes the geometry of the input seismic dataset. This grid defines inline/crossline spacing and survey orientation for the input data, which may differ from the output MF bin grid.
Connect to the trace header collection specifically prepared for MultiFocusing geometry extraction. This collection contains the bin associations and CMP coordinates needed by the MF engine to organize traces into the CO-MF geometry. It must have been populated by a prior binning step. If this collection is empty, the engine will report an error and refuse to run.
Connect to the MF parametrization table item, which controls the search space for MultiFocusing parameters. This table defines the quantized velocity, aperture, and curvature ranges over which the engine searches, along with time tables, minimum fold requirements, dominant frequency, and Fresnel zone size. It is created and configured by the MF Tables module and must be connected before execution.
This table also supports acquisitions composed of multiple survey patches with different nominal fold coverage, where a spatially varying Fresnel Zone factor can be applied to ensure a consistent effective fold across all areas.
Optional. Connect to a velocity constraint picking item from another procedure (for example, a stacking velocity model). When provided, the MF parameter search is constrained to move within the defined velocity corridor rather than searching the full velocity space, which both improves accuracy and reduces computation time.
Optional. Connect to a 2D map (matrix) item containing a spatially varying near-surface velocity (V0) used by the MF engine as the initial reference velocity at each bin location. A spatially varying V0 is important in areas with significant near-surface velocity heterogeneity.
Optional. Connect to a 2D map item representing a spatially varying Fresnel Zone factor. This factor scales the Fresnel Zone size used during geometry extraction on a bin-by-bin basis, which is useful when the survey contains areas of varying acquisition density or nominal fold. When not connected, the Fresnel Zone factor from the MF parametrization table is used uniformly across the survey.
Connect to the on-disk trace vector (TVOD) file handle for reading seismic data directly from disk. This is an alternative to the SEG-Y data handle and trace headers combination. Visible only when Use trace vector on disk is enabled.
When enabled, seismic data is read directly from an on-disk TVOD file handle instead of from a SEG-Y file. This mode is suitable for very large datasets that cannot be efficiently accessed via the SEG-Y interface, or when data has been pre-sorted to disk in a format optimized for bulk random access. Default: disabled.
When using the on-disk trace vector mode, this parameter controls how many traces are read from disk in a single I/O batch. Larger values reduce the number of disk read operations and can improve throughput on fast storage systems, but require more memory. Visible only when Use trace vector on disk is enabled. Default: 200000 traces.
Optional. Path to a velocity corridor file (.corr) that constrains the MF parameter search. When specified, the engine only searches for MF solutions within the velocity bounds defined in the file. This is the recommended approach in areas with a reliable velocity model, as it significantly narrows the search space and reduces computation time. If left empty, the engine searches across the full velocity range defined by the MF parametrization table.
The spatial aperture used to select traces for each bin during the Common-Offset MF calculation. Its physical meaning depends on the CO aperture mode setting: in XY aperture mode this is a lateral distance in meters defining a circular neighborhood around the bin center; in Offset aperture mode it limits the offset range of included traces. Increase this value if the fold within the nominal aperture is too low. Default: 250 m. Minimum: 1 m.
Selects how the CO Aperture value is interpreted when gathering traces for each bin:
XY aperture — selects all traces whose midpoint falls within the specified lateral radius (in meters) of the bin center, regardless of offset class. This is the standard mode for most 3D surveys.
Offset aperture — selects traces by source-receiver offset range instead of spatial proximity. Use this mode when offset-class selection is more meaningful than proximity-based selection, for example in surveys with irregular geometry or narrow azimuth acquisition. Default: XY aperture.
When enabled, the CO-MF engine emulates Zero-Offset (ZO) focusing conditions during the parameter search. This can be useful for QC purposes or when testing how ZO-MF and CO-MF results compare on the same dataset. Default: disabled.
Path to the output 3D MF database file (.kdb). The engine writes the computed MultiFocusing parameters and associated time-corrected events for each bin into this binary database. This file is the primary output of the engine and is required as input for all subsequent Data Enhance (CO-MF) steps. If a database with the same name already exists and its stored parameters do not match the current setup, the engine will prompt whether to continue appending to it, rewrite it, or cancel.
Connect to the bin grid item that defines the output MF geometry — the grid onto which the computed MF parameters will be registered. This grid sets the inline/crossline spacing and orientation for the output KDB database and for subsequent data enhancement. It may differ from the input data bin grid, for example when outputting to a regularized or denser grid than the input acquisition.
This group controls the spatial decimation for storing velocity analysis results in the KDB database. Storing velocity data for every bin can produce very large files; these parameters allow thinning the stored velocity grid to only every N-th inline and crossline.
The step (decimation factor) along inlines for saving velocity analysis data to the KDB database. A value of 5 means velocity information is stored for every 5th inline. This does not affect the quality of the MF parameter search — it only controls the density of the stored velocity QC data. Default: 5. Minimum: 1.
The step (decimation factor) along crosslines for saving velocity analysis data to the KDB database. A value of 5 means velocity information is stored for every 5th crossline. Default: 5. Minimum: 1.
This group defines the spatial and temporal extent of the MF parameter search. Use it to limit computation to a specific portion of the survey (for example, to test on a small patch, or to complete a partially processed survey). The interactive location map (vista panel) displays the planned calculation area, already-completed bins, and bins marked for reprocessing.
When enabled, the polygon defined by the Polygon points vector input is used to mark bins inside the polygon for recalculation, even if they were previously computed. This is useful for selectively reprocessing a sub-area of the survey after parameter adjustments. Default: enabled.
Optional. Connect to a point vector item defining a polygon boundary. When Recalculate polygon mode is enabled, only bins whose centers fall inside this polygon will be flagged for recalculation. This provides a flexible way to target specific survey areas for reprocessing without changing the inline/crossline range limits.
The first inline number to include in the calculation. Set to -1 (default) to start from the first available inline in the survey. Increasing this value restricts the computation to a subregion, which is useful for testing or incremental processing.
The last inline number to include in the calculation. Set to -1 (default) to process through the last available inline. Together with First inline number, this defines the inline extent of the calculation area.
The first crossline number to include in the calculation. Set to -1 (default) to start from the first available crossline. Combine with the inline limits to define a rectangular calculation area within the full survey.
The last crossline number to include in the calculation. Set to -1 (default) to process through the last available crossline. Default: -1.
Controls how the set of bins to calculate is determined:
Constant grid — bins are selected on a uniform inline/crossline grid defined by the step parameters. This is the standard mode for full-survey processing.
Selection by min offset — bins are selected based on minimum offset criteria. This mode activates the Min offset calculation, Merge distance, and Desired min distance parameters and allows the calculation area to adapt to the actual coverage of the dataset rather than using a fixed step grid. Default: Constant grid.
The maximum allowed minimum offset (in meters) for a bin to be included in the calculation when using the Selection by min offset mode. Bins whose minimum available source-receiver offset exceeds this threshold are excluded. Default: 100000 m (effectively no limit).
When using the Selection by min offset mode, this distance (in meters) controls the spatial merging of nearby candidate bins. Bins whose centers are closer than this distance to each other may be merged to avoid redundant computation in densely covered areas. Default: 0 m (no merging).
The target minimum spacing (in meters) between selected bins when using the Selection by min offset mode. This ensures the selected bins are not excessively densely spaced. Default: 1000 m.
Determines which bins are processed when Constant grid selection is active and the step values are greater than 1:
Logical AND — a bin is calculated only if it lies on both the inline step grid and the crossline step grid simultaneously (i.e., at the intersections of the sparse grid). This produces a regularly spaced 2D grid of calculated bins.
Logical OR — a bin is calculated if it lies on the inline step grid or the crossline step grid (i.e., along both inline and crossline swaths). This produces a cross-shaped or grid-line pattern of calculated bins, which is useful for rapid QC passes. Default: Logical AND.
The inline decimation step for the calculation grid when using Constant grid selection. A value of 1 calculates every inline; a value of 2 calculates every other inline, and so on. Use larger values for rapid pilot tests and reduce to 1 for production processing. Default: 1.
The crossline decimation step for the calculation grid when using Constant grid selection. A value of 1 calculates every crossline; larger values thin the grid proportionally. Default: 1.
The start of the time window (in seconds) within which the MF parameter search is performed. Events shallower than this time are ignored. Set this to skip very shallow, low-quality, or noisy data (e.g., direct wave and refraction arrivals). Default: 0 s.
The end of the time window (in seconds) for the MF parameter search. Events deeper than this time are not processed. Set this to the deepest horizon of interest to avoid wasting computation on noise-dominated deep data. Default: 6 s.
When enabled, bins that have already been calculated in a previous run but are interactively flagged for reprocessing (shown in red on the location map) will be recalculated. When disabled, all previously computed bins are skipped regardless of their selection status. Use this option to efficiently update specific bins after reviewing QC results, without rerunning the entire survey. Default: disabled.
This group contains parameters for special-case workflows. Leave at default values unless you have a specific reason to change them.
When enabled, the engine skips the consistency check that compares the stored KDB database headers against the current module parameters before opening the database. This allows appending to an existing database even if some parameters have changed. Use with caution — enabling this option can result in mixed or inconsistent results in the KDB file if the parametrization has changed significantly. Default: disabled.
The maximum number of traces allowed per offset class when populating the CO gather for each bin. If the actual fold exceeds this limit, traces are thinned to this number before MF parameter search. This prevents excessive computation time in areas of very high nominal fold (e.g., in the center of a 3D survey). Default: 10000. Minimum: 1.
A bin-point vector item representing the output geometry of the MF calculation — i.e., the list of bins for which MF parameters have been computed. Connect this output to downstream processing steps that require knowledge of the MF-computed bin locations.
Refreshes the interactive location map to reflect the current calculation area settings (inline/crossline limits and steps) and shows which bins are planned for calculation, which are already computed, and which are flagged for reprocessing. Run this action after changing any calculation area parameters to update the visual display before executing.
Resets the calculation flags for all bins that fall inside the polygon defined by the Polygon points vector input item. After running this action, those bins will be treated as not yet calculated and will be included in the next execution run.
Creates a spatial mask from the bins currently selected on the interactive location map (the bins shown in the Picked Bins overlay). The mask is stored internally and displayed in the Mask bins overlay. This mask is required before running the Delete by mask action.
Permanently deletes computed MF data from the KDB database for all bins that are within a user-specified distance of the current mask. The engine will ask for confirmation and then prompt for the search distance (in meters). This action is irreversible — deleted bin data cannot be recovered without rerunning the engine. Use this action to remove erroneous or corrupted bin results from the database before selectively recomputing them.