|
<< Click to Display Table of Contents >> Navigation: MultiFocusing > Engine - 3D CO-MF Constrain |
Engine - 3D CO-MF Constrain is a deprecated 3D Multi-Focusing (MF) search engine that operates in Common-Offset (CO) mode with an additional constraint gather. It scans each CO virtual bin within the survey area and searches for the optimal Multi-Focusing wavefront parameters — velocity, emergence angle, and radius of curvature — using a constrained semblance optimizer. Unlike the standard 3D CO-MF engine, this module supplies a secondary low-offset (minimum-offset) gather as an amplitude and phase constraint to stabilize the parameter search in bins with sparse or low-quality far-offset data.
The module writes all computed MF parameters directly to a 3D CO storage database (.kdb file). It supports distributed execution across multiple CPU cores and GPU devices, and can resume a partially completed run without reprocessing already-calculated bins. The search space is defined by the MF Tables Parametrization item, and may optionally be constrained to a velocity corridor from a picking file or a connected GMFPickingItem. A Fresnel zone aperture scaling map (FZFactorMatrix) and a spatial V0 map can also be provided to improve parameter quality in areas of variable near-surface velocity.
The calculation area can be restricted to a sub-volume of the survey by setting inline and crossline range limits, and bins can be selected either on a constant sparse grid (every N-th inline and crossline) or automatically by minimum-offset coverage using the "Selection by min offset" mode.
An optional velocity corridor picking item (G3DPickingItem) that constrains the MF parameter search space. When connected, the engine only evaluates wavefront parameters that fall within the corridor defined by this picking. This prevents the optimizer from returning physically unreasonable solutions in noisy or low-fold areas. If neither this item nor a picking file is connected, the full parameter table is searched without constraint.
An optional spatial map (GMatrixItem) of Fresnel zone aperture scaling factors. When connected, the aperture used to extract traces for each CO bin is multiplied by the local factor read from this map. This allows the effective aperture to vary spatially — for example, tightening it in structurally complex areas and widening it where the wavefield is smooth. If not connected, a uniform aperture defined by the CO Aperture parameter is used throughout the survey.
The SEG-Y file handle providing access to the pre-stack seismic amplitude data. The engine reads trace gathers on demand during processing. The data must be sorted as 3D CMP gathers, and the sorted headers index must match the sort order of this file.
The sorted trace-headers index for the input seismic data. This index must describe a 3D CMP geometry; the module validates this requirement at startup and will report an error if the index is not 3D CMP-sorted. It is used to efficiently locate all traces belonging to any given bin during the MF parameter search.
The bin grid describing the geometry of the input seismic data. This defines how inline and crossline numbers map to spatial XY coordinates and is used when extracting aperture-windowed gathers around each virtual CO bin centre. It must match the geometry embedded in the SEG-Y data.
The virtual CO geometry trace-header collection, typically produced by the "Virtual geometry COMF 3D" module. Each entry describes a virtual CO trace: its central bin, offset class, and associated source-receiver geometry. This collection defines the set of bins that the engine will attempt to process, and must have a non-empty bin set (the module checks this and reports an error if the collection is empty).
The Multi-Focusing parameter-table quantization item, produced by the "Create MF tables" module. It defines the discretized search space: the velocity table, angle table, time table, correlation window length, selection window length, minimum fold, main frequency, and minimum Fresnel zone. All engine search parameters — including the inline and crossline aperture used for gather extraction — are taken from this item. This input is mandatory.
An optional spatial map (GMatrixItem) of the near-surface reference velocity V0. When connected, the engine uses the spatially varying V0 value at each bin location to parameterize the Multi-Focusing operator, improving accuracy in areas of strong near-surface velocity variations. If not connected, the V0 value encoded in the MF Tables Parametrization is used uniformly across the survey.
Path to an optional velocity corridor file (.corr format) that constrains the MF parameter search. If supplied, the engine loads the corridor picking before execution and uses it to restrict the wavefront parameter search to physically plausible values at each time sample and bin location. This is an alternative to connecting a live GMFPickingItem — either one or the other may be used. If both are specified, the file takes precedence. Leave blank to run an unconstrained search.
The maximum offset aperture (in metres) used to select real traces around each virtual CO bin centre when building the constrained MF gather. Default: 250 m. Only traces whose midpoint falls within this radius are included in the semblance computation. Increasing this value brings in more traces and may improve the signal-to-noise ratio of the semblance panel, but increases computation time. Set this value to match the expected spatial coherency of the reflections in your data.
This container groups parameters related to the 3D MF storage database.
Path to the output 3D CO Multi-Focusing storage database (.kdb). This binary database stores the computed MF wavefront parameters (velocity, angle, curvature radius, semblance) for every processed bin and time sample. The database is also used for restart: if the file already exists and its geometry matches the current run, only uncalculated bins are processed. If the geometry does not match, the module prompts whether to overwrite the existing database. This parameter is mandatory.
This container groups the output geometry definition for the MF database.
The bin grid defining the geometry of the output MF database. This may differ from the input seismic bin grid — for example, the output grid may be coarser or cover only a sub-area of the survey. The engine uses this grid to assign computed MF parameters to database locations and to determine the spatial coverage of the "Calculation area" vista map. This input is mandatory for execution.
This container groups parameters that control how densely MF parameters are saved into the database for subsequent velocity analysis.
The inline sub-sampling stride for writing semblance and parameter data into the storage database. Default: 5. A value of 5 means that full semblance panels are written for every 5th inline; intermediate inlines store only the picked MF parameters, not the full semblance cube. Reduce this value to retain more detailed velocity analysis data at the cost of larger database files. Minimum value: 1 (save every inline).
The crossline sub-sampling stride for writing semblance data into the storage database. Default: 5. Works identically to "Each inLine" but along the crossline direction. Together, these two parameters define the density of the stored semblance data grid available for interactive velocity analysis after the run completes. Minimum value: 1.
This container groups all parameters that define which bins within the output bin grid are included in the MF parameter search. Use these settings to limit processing to a specific area of the 3D survey, to perform test runs, or to restart processing for a sub-volume after parameter adjustments.
The first inline number to include in the calculation. Default: -1 (no limit — use the first inline of the output bin grid). Set to a specific inline number to restrict processing to a sub-area. The effective start is the maximum of this value and the first inline of the output bin grid.
The last inline number to include in the calculation. Default: -1 (no limit — use the last inline of the output bin grid). Set to a specific inline number to stop processing before the end of the survey. The effective end is the minimum of this value and the last inline of the output bin grid.
The first crossline number to include in the calculation. Default: -1 (no limit — use the first crossline of the output bin grid). Used together with "Last crossline number" to define a crossline sub-range for processing.
The last crossline number to include in the calculation. Default: -1 (no limit — use the last crossline of the output bin grid). Set to a specific crossline number to stop processing before the edge of the survey crossline range.
Controls how bins are selected for processing within the defined inline/crossline range. Two options are available:
Constant grid (default): Bins are selected on a regular sparse grid defined by the "Step along inLine number" and "Step along xLine number" parameters, with the "Calculation rule" controlling whether a bin must satisfy both steps simultaneously (Logical AND) or either step (Logical OR).
Selection by min offset: Bins are selected based on the minimum-offset coverage of the input seismic data. A subset of representative low-offset bins is identified automatically, and all CO bins are mapped to their nearest such representative. This mode reveals the three additional parameters "Min offset calculation", "Merge distance for min offset calculation", and "Desired min distance", and forces the inline/crossline step to 1.
Active only when "Selection type" is set to "Selection by min offset". Specifies the maximum minimum-offset value (in metres) that a bin must have to be considered as a constrain reference point. Default: 100000 m. In practice, set this to a geologically meaningful near-offset threshold — bins with a minimum offset below this value are used as constraint gathers for their neighbours. Reducing this value selects only the best-covered near-offset locations as reference points.
Active only when "Selection type" is set to "Selection by min offset". Specifies a merging distance (in metres) used to consolidate nearby reference bins into a single representative point. Default: 0 m (no merging). Increase this value to reduce the number of distinct constraint reference bins when the low-offset coverage is spatially dense and redundant.
Active only when "Selection type" is set to "Selection by min offset". Sets the desired spatial spacing (in metres) between selected constraint reference bins. Default: 1000 m. The selector attempts to place reference points no closer than this distance to one another. Adjust this value to balance the spatial density of constraint information against computation time.
Active when "Selection type" is "Constant grid". Controls the logical relationship between the inline step and the crossline step when selecting bins. Default: Logical AND.
Logical AND: A bin is processed only if it falls on both the inline grid step and the crossline grid step simultaneously. This produces a sparse rectangular grid of processed bins.
Logical OR: A bin is processed if it falls on the inline step OR the crossline step. This produces a cross-hatch pattern covering more bins than Logical AND and is useful for QC runs that need denser coverage along both directions without processing every bin.
Active when "Selection type" is "Constant grid". The inline sub-sampling step: only every N-th inline (relative to the first inline of the calculation range) is included in the calculation. Default: 1 (every inline). Increase to reduce computation time during initial parameter testing. Minimum value: 1.
Active when "Selection type" is "Constant grid". The crossline sub-sampling step: only every N-th crossline is included in the calculation. Default: 1 (every crossline). Use together with "Step along inLine number" to define a coarse test grid before running the full survey. Minimum value: 1.
The start time (in seconds) of the MF parameter search window. Default: 0 s. The engine computes MF parameters only within the time interval [First Time, Last Time]. This allows processing to focus on a target geological interval and reduces computation time when deep reflectors are not of interest.
The end time (in seconds) of the MF parameter search window. Default: 6 s. Set this to the two-way travel time of the deepest target reflector to avoid unnecessary computation on noise-dominated late times.
This container groups advanced options that control how the module handles geometry conflicts with an existing database.
When enabled, the module skips the database header compatibility check and does not prompt the user if an existing .kdb file has different geometry parameters. Default: off (the module asks for user confirmation before overwriting an incompatible database). Enable this option in automated or batch workflows where a user prompt cannot be answered interactively, or when you intentionally want to overwrite an existing database without confirmation.
Refreshes the "Calculation area" vista map overlay, which displays the set of bins that will be processed given the current inline/crossline range, step, and selection-type settings. Run this action after changing any calculation area parameter to preview which bins are selected before starting the full computation. The map shows all candidate bins as points; in "Selection by min offset" mode it additionally displays the constraint reference bins as a separate overlay ("Picked Bins").