Description

This is the interactive velocity-building tool for 2D Kirchhoff Prestack Time Migration (KPTM). It reads the migration storage database produced by the companion Kirchhoff PSTM Engine step, which contains Common Image Gathers migrated with a full range of trial velocities (a velocity scan) at every bin along the line, and turns that scan into a working migration velocity field.

The module offers two velocity-picking workflows. In the default semi-automatic mode it searches the semblance (image coherency) volume for a sparse grid of the strongest events, then interpolates those automatic picks into a smooth, spatially consistent velocity field across the line. In the legacy manual mode you place velocity control points by hand on the semblance panel and on the Common Image Gathers, and the field is built from those points using Kriging, Triangulation, or ABOS interpolation. In both modes the module builds a final velocity field and the corresponding migrated stack image, which you can QC directly against the semblance and CIG displays before handing the velocity field off to the next processing or depth-conversion step.

Custom actions

This module exposes the following toolbar/menu actions in addition to its normal execution.

Recreate Migrations

Forces a full recompute of the stacked migration panels and of the semblance / velocity / velocity-factor volumes for every trial velocity and every bin, discarding any previously cached results. Use it after changing the Mute, AGC, or Hilbert settings, or whenever the underlying migration storage has been updated, since normal execution reuses previously computed results when available.

Calculate sparse max semblance

Runs the first step of the semi-automatic (new mode) workflow: scans the full semblance volume, using the Sparsing window settings, and keeps a sparse grid of the strongest local coherency maxima as automatic velocity picks.

Calculate interpolated velocity

Runs the second step of the semi-automatic workflow: interpolates the sparse grid of automatic picks into continuous semblance, velocity, and velocity-factor volumes covering every trace and time sample, using the selected Interpolation Method (new mode).

Calculate interpolated stack

Runs the final step of the semi-automatic workflow: builds the final velocity field from the interpolated volumes and the matching migrated stack image, and refreshes the Image Stack, Final Velocity, and Final Velocity Smooth outputs.

Clear all picking

Deletes all manually placed velocity control points (used by the legacy manual-picking workflow), after a confirmation prompt.

Load picking

Opens a previously saved velocity-picking file (*.velpick) and loads its manual velocity control points into the current session.

Save picking

Saves the current manual velocity picking to a picking file (*.velpick) that can be reloaded later or shared between sessions or lines.

Convert picking

Converts an older picking-corridor file (*.corr) into the current picking file format, writing the converted result next to the original file with a "_newPick" suffix.

Clear mute

Deletes all mute picking points (used when "Use Mute" is enabled), after a confirmation prompt.

Load mute

Loads mute picking points from a corridor file (*.corr) into the current session.

Input data

Storage file

The KPTM migration storage database (.kdb) produced by the companion 2D Kirchhoff Prestack Time Migration Engine step. For every bin along the line it holds the Common Image Gathers migrated with the full range of trial velocity (scan) factors that this module scans, semblance-analyzes, and interpolates to build the working velocity model. This is a read-only input; select the storage file for the line you want to pick velocities on.

GMutePickingItem

Stores the interactive mute picks used to exclude unwanted energy, such as noise trains, direct arrivals, or multiples, from each Common Image Gather before stacking and semblance calculation, whenever "Use Mute" is enabled. You edit these picks directly on the CIG display with the Manual Mute drawing tools; they are saved and reloaded together with the project.

GVerticalVelocityPickingItem

Companion storage for vertical (interval) velocity picking curves associated with this line, allowing velocity information built or reviewed here to be shared with other vertical-velocity-picking tools in g-Platform. It is maintained automatically as part of the project and does not normally need to be set manually.

Parameters

Mode

Selects which velocity-building workflow the module uses. When enabled, the module works in the semi-automatic grid mode: it automatically searches the semblance volume for a sparse grid of the strongest coherent events and interpolates a smooth velocity field from those automatic picks (the "New Mode" parameter groups below apply in this mode). When disabled, the module switches to the legacy manual-picking workflow, where you place velocity control points by hand on the semblance panel and Common Image Gathers, and the field is built from those points with Kriging, Triangulation, ABOS, or NearestSmooth interpolation (the "Interpolation method (old mode)" and "Kriging Parameters" settings apply in this mode).

Default: On (semi-automatic grid mode). Turn this off only for projects that rely on the legacy manual-picking workflow, since most of the parameter groups below are specific to one mode or the other.

Velocity number

Selects which single velocity-scan panel, out of the full range of trial velocities computed by the Engine step, is currently displayed in the Image Stack and Current Velocity Factor viewers. Numbering starts at 1. Step through this value to visually compare migrated panels at different stacking velocities before relying on the automatic or manual picking to select the best one.

Default: not set. Valid values run from 1 up to the number of velocity scans stored in the migration database (or up to "Number of velocities" under the Patch group, if Patch parameters are in use).

Interpolation method (old mode)

Only used when Mode is set to the legacy manual-picking workflow. Determines how your manually placed velocity control points are turned into a continuous velocity-factor surface: Kriging performs a geostatistical interpolation weighted by point spacing, using the Kriging Radius and Kriging Number of Points below; Triangulation builds a triangulated surface directly through the picks; ABOS produces a smoothed, grid-based surface from the picks; NearestSmooth assigns each grid cell the value of its nearest pick and then applies lateral smoothing controlled by Smoothing Parameter X and Smoothing Parameter Y.

Default: NearestSmooth. Prefer Kriging for sparse, irregularly spaced picks that benefit from statistical weighting; prefer Triangulation for dense, evenly spaced picks; use NearestSmooth or ABOS when you want a smoother field that is less sensitive to individual outlier picks.

TaperWin

Length of the cosine taper applied at the edge of the mute whenever "Use Mute" is enabled, so muted samples fade out smoothly instead of being cut off abruptly. This avoids sharp discontinuities (ringing) in the semblance and stacked panels near mute boundaries.

Default: 0 ms (no taper). Increase this if you notice sharp artifacts in the semblance or stack near mute edges.

Debug / technical options

Recalculate semblance

Technical option grouped with the SEG-Y export toggle below, related to forcing the semblance and velocity-scan volumes to be rebuilt rather than reused. For normal work, use the "Recreate Migrations" custom action described above to force a full rebuild; leave this checkbox at its default for everyday picking.

Default: Off.

Create SEG-Y Files To D://Temp

When enabled, every intermediate volume the module computes (maximum semblance, velocity, velocity-factor, sparse-grid, interpolated, and final stack panels) is additionally written out as SEG-Y files to D:\Temp on the processing machine. This is a diagnostic aid for troubleshooting picking or interpolation results with external tools; it is not required for normal production work and adds extra processing time and disk usage.

Default: Off. Enable temporarily only when you need to inspect intermediate results outside g-Platform.

Kriging parameters (old mode)

Kriging number of points

Maximum number of nearby manual picks used to estimate the velocity factor at each grid position when the Interpolation method (old mode) is set to Kriging. Increase for a smoother, more regional estimate that averages more neighboring picks; decrease to let the field follow local picks more tightly.

Default: 30 points.

Kriging radius

Search radius, in meters along the line, within which picks are considered by the Kriging interpolator when estimating each grid position. Increase for widely spaced picks so enough points are found around each location; decrease for densely picked lines to keep each estimate local.

Default: 50 m.

Smoothing parameter X

Lateral (along-line) smoothing distance applied when building the velocity-factor surface with the NearestSmooth interpolation method. Larger values produce a laterally smoother velocity field; smaller values preserve more local detail from individual picks.

Default: 50 m.

Smoothing parameter Y

Vertical (time-axis) smoothing distance applied together with Smoothing parameter X during NearestSmooth interpolation. Increase to blend velocity picks over a longer time window; decrease to keep the field more responsive to velocity changes with time.

Default: 50 m.

Visualization

Datum

Reference elevation used whenever "Shift to Datum" (or its light variant) is enabled. The semblance, CIG, and final velocity/stack displays are statically shifted to this datum so that panels from bins at different surface elevations line up in the same time frame for comparison and picking.

Default: 0 m. Set this to the project's standard processing datum.

ShiftToDatum

When enabled, the module statically shifts semblance, CIG, and the final velocity/stack outputs to the elevation set in Datum, compensating for surface topography so that picks made across bins at different elevations remain time-consistent.

Default: Off. Turn on for 2D lines with significant topographic relief; leave off for flat-datum marine data or data that is already statically corrected.

ShiftToDatumLight

Lighter-weight variant of the datum shift, applied for quick on-screen preview updates while interactively picking rather than for the full computed output. Keep this on for responsive interaction; it does not replace the full "Shift to Datum" setting for the final computed outputs.

Default: On.

Load CIG

Controls whether the Common Image Gather for the currently selected bin is loaded and displayed while interactively QC'ing or picking. Turn this off to speed up interaction on very large storage files when you only need the semblance and velocity-scan panels and do not need to inspect individual CIGs.

Default: On.

Mute

Use Mute

Applies the mute picking defined in the "GMutePickingItem" input, edited interactively with the Manual Mute tools, to every Common Image Gather before it is stacked and before semblance is calculated. Enable this to exclude noisy shallow or far-offset energy, such as direct arrivals, refractions, or noise trains, from the velocity scan and stack; this is especially useful on data with strong near-offset noise.

Default: Off.

Patch

Use patch params

When enabled, overrides the base-velocity index and the number of velocity scans normally read from the migration storage file with the two values below, letting you rebuild or "patch" the velocity picking using a different velocity range than the one originally computed by the migration Engine step.

Default: Off. Leave off to use the full velocity range stored in the migration database.

Index of base velocity

Only used when "Use patch params" is enabled. Selects, by position starting at 1, which of the stored velocity scans is treated as the reference (base) velocity for computing velocity factors.

Default: 5.

Number of velocities

Only used when "Use patch params" is enabled. Sets how many consecutive velocity scans, starting at the base velocity index above, are loaded and used for semblance and stack calculation instead of the full set stored in the migration file. Use a smaller number to quickly test picking on a reduced velocity range before committing to the full scan.

Default: 5.

Sparsing (new mode)

Sparse window size X

Only used in the semi-automatic (new mode) workflow. Width, in traces, of the search window used when scanning the semblance volume for local maxima during the "Calculate sparse max semblance" step. Increase to reduce the density of automatic picks along the line, which speeds up interpolation and produces a smoother, less locally driven velocity field. This value is ignored (forced to 1) when the Interpolation method (new mode) below is set to "Linear By Each Trace".

Default: 1 trace. Minimum value: 1 trace.

Sparse window size Y

Height, in time, of the search window used together with Sparse window size X when scanning for local semblance maxima. Increase to look for the single strongest event over a larger time window, giving fewer but more confident automatic picks per trace; decrease to allow more closely spaced picks that better follow rapid velocity changes with time.

Default: 0.025 ms. Minimum value: 0.001 ms.

Interpolation method (new mode)

Only used in the semi-automatic (new mode) workflow. Determines how the sparse grid of automatically detected maximum-semblance picks is turned into continuous semblance, velocity, and velocity-factor volumes covering every trace and time sample. Triangulation builds a triangulated surface through the sparse picks, pinning the section corners to the nearest picks. Distance performs a distance-weighted interpolation. Smooth and Smooth By Max Semblance blend neighboring picks with a moving-average style smoothing, with the latter also weighting by semblance strength. Linear By Each Trace instead interpolates each trace independently between its own picked control times using linear interpolation, and also forces Sparse window size X to 1 trace.

Default: Linear By Each Trace, the most predictable, per-trace option. Switch to Triangulation or one of the Smooth methods when you want the module to blend information laterally between neighboring traces and bins.

Smooth

Ignore value

Value treated as "no data" and intended to be excluded from the final-velocity smoothing pass, so that empty or invalid grid cells do not bias the smoothed result.

Default: 0.

Smooth grid X

Horizontal (along-line) smoothing window, in traces, used by the weighted velocity smoothing filter that builds the "Final Velocity Smooth" output from the raw Final Velocity field. Increase for a laterally smoother final velocity model, useful for regional velocity trends; decrease to preserve more local lateral velocity detail.

Default: 20 traces.

Smooth grid Y

Vertical (time-axis) smoothing window, in samples, used together with Smooth grid X by the same weighted velocity smoothing filter. Increase for a smoother velocity model with depth or time; decrease to retain sharper velocity contrasts, for example across a strong impedance boundary.

Default: 20 samples.

Use adaptive window

Lets the velocity smoothing filter adapt its window size automatically instead of always using the fixed Smooth grid X / Smooth grid Y sizes above.

Default: Off.

Max iterations count

Maximum number of smoothing passes applied when building the smoothed final velocity field. Increase for a more thoroughly smoothed result; decrease to keep processing time down and preserve more of the original picked detail.

Default: 5.

Smooth on datum

When enabled, the final velocity field is first statically shifted to the datum elevation set under Visualization, smoothed there, and then shifted back to topography, so smoothing operates on a flattened section rather than across topography-distorted time. Enable this on lines with significant relief to avoid smoothing velocities across artificial time shifts caused by elevation changes.

Default: Off.

New Mode - Semblance

SemblanceThreshold

Minimum semblance (image coherency) value, on a 0 to 1 scale, that an automatically detected local maximum must reach to be kept as a valid velocity pick in the semi-automatic workflow. Raise this to keep only strong, reliable coherency picks and reduce noise-driven mis-picks; lower it if valid picks are being rejected on noisy or low-fold data.

Default: 0.1. Valid range: 0 to 1.

MinTime

Earliest time in the section considered when searching for automatic semblance maxima. Raise this to exclude shallow time gates that are often noisy or contaminated by multiples from the automatic picking.

Default: 0 ms. Minimum value: 0 ms.

MaxTime

Latest time in the section considered when searching for automatic semblance maxima. Set this to, or just beyond, the deepest reflector of interest; automatic picks below this time are discarded, which also speeds up the search.

Default: 1000 ms. Minimum value: 0.001 ms.

MinVelocityFactor

Lower bound on the velocity factor, the ratio of trial velocity to the base/reference velocity, accepted for automatic picks. Raise this to reject unrealistically slow automatic picks.

Default: 0. Minimum value: 0.

MaxVelocityFactor

Upper bound on the velocity factor accepted for automatic picks. Lower this to reject unrealistically fast automatic picks and keep the resulting velocity field geologically reasonable.

Default: 100. Minimum value: 0.

SemblanceWindowSize

Length of the time gate used to compute and smooth the semblance (coherency) measure at each time sample. Increase for a more stable, less noisy semblance measurement at the cost of vertical resolution; decrease to better resolve closely spaced velocity changes.

Default: 0.020 ms. Minimum value: 0.001 ms.

SemblanceNormalizationOfFluctuation

When enabled, the semblance calculation is normalized to reduce amplitude-driven fluctuations, giving a more consistent coherency measurement across the line regardless of local amplitude variations. Leave this enabled for most data; disable it only for diagnostic comparison against the raw, unnormalized semblance.

Default: On.

New Mode - CIG/Semblance

Use Hilbert For CIG

When enabled, the Common Image Gather display and semblance calculation use the Hilbert-transformed (envelope) version of the data, which reduces phase-related mis-ties when comparing coherency across offsets and trial velocities.

Default: On.

Use AGC For CIG

Applies Automatic Gain Control to the Common Image Gather before semblance calculation and display, boosting weak, deeper energy so it is easier to see and pick on the semblance panel. Enable this on data with strong amplitude decay with time or offset so that deep reflectors remain visible for picking.

Default: Off.

AGC Window Time For CIG

Length of the sliding time window used by the AGC when "Use AGC For CIG" is enabled. Shorter windows react faster to amplitude changes for a more aggressive gain; longer windows produce a smoother, more stable gain.

Default: 0.100 ms. Minimum value: 0.001 ms.

Output data

Image Stack

The final migrated (PSTM) stack section for the line, built by blending the pre-migrated velocity-scan panels according to the final interpolated velocity field. This is the production stack output for the 2D line once velocity picking is complete, and it is also the panel displayed for interactive QC while picking.

Final Velocity

The final migration velocity field, in the same physical units as the base velocity (m/s), computed as the base/reference velocity multiplied by the picked and interpolated velocity factor at every bin and time sample, honoring the Datum shift settings if enabled.

Final Velocity Smooth

A laterally and vertically smoothed version of the Final Velocity field, built using the Smooth grid X / Smooth grid Y and related Smoothing settings. It is suitable for use as a well-behaved velocity model for depth conversion or further migration passes, without the small-scale artifacts that can come directly from individual velocity picks.

GGatherFactorMap

The interpolated velocity factor field itself, the ratio of the picked velocity to the base/reference velocity at each bin and time sample, optionally shifted to topography. This is useful for QC of the spatial consistency of the velocity picking independent of the absolute velocity values.

References

Yilmaz, O., 1987, Seismic data processing: Society of Exploration Geophysicists.

Al-Yahya, K. M., 1989, Velocity analysis by iterative profile migration: Geophysics, 54, 718-729.