Cross-plot Tool in OpendTect
Farrukh Qayyum
Introduction
This presentation is setup to introduce the
Cross-plot tool available in OpendTect.
This will familiarize you with
How to launch the cross-plot tools,
How to use the available options
Cross-plot manipulations
Examples
Outline
Purpose and Benefits
How To ...
Menus/short-cuts
Attributes vs Attributes cross-plotting window
Attributes vs wells cross-plotting window
Attribute data window
Main cross-plot window (settings and manipulation)
Cross-plot Examples
Purpose and Benefits
The cross-plot tool in OpendTect supports:
Attribute Vs Attribute Cross-plot
Attribute Vs Log Cross-plot
Log Vs Log Cross-plot
Cross-plots can be used for
Object detection (supervised-Neural Networks, logical prediction, body
extraction etc).
Well properties prediction
Linear/non-linear prediction
Finding a relationship between multi-attributes (good or bad
attributes) or between attribute/logs for Meta-Attribute analysis
On the fly interaction between active scene and the cross-plot
etc.
Outline
Purpose and Benefits
How To ...
Menus/short-cuts
Attributes vs Attributes cross-plotting window
Attributes vs wells cross-plotting window
Attribute data window
Main cross-plot window (settings and manipulation)
Cross-plot Examples
How To
Menu/Shortcuts
Launch Attribute Vs Well Logs Cross-plotting window
Attributes vs Well
logs Crossplot Icon
How To
Menu/Shortcuts
Launch Attribute Vs Attributes Cross-plotting window
Attributes Cross-plot
icon
Attribute-Well Cross-plotting window
How To
Attributes Well logs cross-plotting window
[1] Multiple selection of input seismic attributes
defined in the attribute set window or stored
volumes
[2] Select multiple wells and corresponding logs
[3] Replicate the logs around the specified radius
(uses survey XY units)
[4] Extraction filter (see next slide)
[5] Limit the data extraction between the markers
Attribute Cross-plotting window
How To
Attributes - Attributes cross-plotting window
[1] Multiple selection of input seismic attributes
defined in the attribute set window or stored
volumes
[2] Extract the scattered data in a volume, along a
surface, between surfaces, within a polygon etc.
[3] Extraction filters:
A user can reduce the amount of
extracted data by applying several
filters. For instance, extraction of
scattered data between two
surfaces[2] but within a polygon.
Attribute Data
How To
Extracted Data - Spreadsheet
Next Slide
These fields are editable !
Toggle XY-coord or Z columns in the
Select X, Y1, Y2 axis Move selected Y-axis left /right
Save data spreadsheet are turned ON/OFF
Remove selected Y-axis Sort selected column
Open stored data Display histogram of the
selected column
Deleted selected rows Display 7.7% of extracted data
Create cross-plot
How to work with Attribute Data window
How To
Save/Open Data - Spreadsheet
A useful option to save time. Store cross-plot data and retrieve later for further analysis
The extracted data can be saved as a simple text file Data stored in OpendTect format can be retrieved
(e.g. for analysis in EXCEL) or as an OpendTect through this button.
format (*.pvds)
Attribute Data: Axis Selection
How To
Axes Selection - Spreadsheet
4-Press it to create a
cross-plot (X, Y1, Y2)
1-Click here and
press 2-Click it and press 3-Click it and press
to add Y1 to add Y2
Y1 Y2
In order to select X, Y1, and Y2 axes, use the following buttons.
First click on a column to highlight it and press X/Y. Y+ refers to add additional Y-axis (Y2).
This will change the selected Y-axis to another attribute/log and will display the corresponded Y-axis on the
cross-plot.
Attribute Data: Remove unwanted rows
How To
Remove Rows - Spreadsheet
Highlight multiple lines and remove the unwanted
(e.g. null values, 0 values etc)
Removes selected rows
Attribute Data: Interactive cross-plot update
How To
Interactive Cross-plot (axes) Update - Spreadsheet
A cross-plot is updated interactively, if
Another column is selected for X or Y
A data (row) is removed
The fields (data) are edited
This will change the selected Y-axis column to another attribute/log column and
it will update the corresponding Y-axis of the cross-plot.
Attribute Data: Histogram
How To
Histograms/Statistics - Spreadsheet
A quick display of histogram of any selected
column in a spread sheet.
QC- the extracted points and check the data
statistics (mean, median, standard derivations
etc)
Cross-plot window
How To
Cross-plot Window
Correlation coefficients
Only 2% of extracted data is displayed here
(Used to avoid issues if the displayed data
exceeds available memory)
Toggles density plot On
If pressed, Y2-axis will be plotted. If not pressed, only Y1 will
be plotted.
Selection of the scattered data can either be made on
Next Slide Y1 or both
Interact mode used to make selection by drawing polygon/rectangle
Pan mode, used to pan the display (if zoomed in)
Displays the selected scattered data (as green picks) in the
active scene. This is the tool for creating objects (picksets/body)
Selection of scattered data (polygon and rectangle). Use in
combination with interact mode and left mouse button dragging.
Select or deselect points selected by
Removes the selected points from the spreadsheet.
Shows selected points in the spread sheet that can also be
removed by using
This is used to make a conditional settings in the selection. For
instance, if a user wants to select data points by avoiding some
values. To learn more go to the Appendix at the end of the
Takes a snapshot of the cross-plot OpendTect manual
Cross-plot settings (see next slide)
Cross-plot Settings
How To
Settings Crossplot Window
Launches the following window
Scaling: Statistics:
Sets the axis (X, Y, Y1) ranges. This tab shows linear regression
statistics of the cross-plot data.
-By clipping certain % of data
(if Use clipping- Yes) -Put in plot (shows correlation
coefficient in the cross-plot)
-By setting manual (if Use
clipping- No) -Check show regression line to
show a regression line in the
cross-plot
User Defined: Density plot*:
Set user defined regression This tab is used to specify a
statistics (Y=a+bX) and display threshold value (no. of points)
the regression line in the cross- to plot a density plot. For
plot instance; if the current number
of points are 11100, and the
minimum is set to 5000, then
whenever the number of points
exceeds 5000, the cross-plot is
displayed as a density plot.
* Useful to avoid a memory
issues
Normal Vs Density Cross-plot
How To
Normal Vs Density Plots Crossplot Window
Toggles Normal plot Toggles Density plot
Density Plot Normal Plot
A density cross-plot is useful to avoid memory issues. Only two axes (X, Y1) are displayed in a density plot.
A normal plot displays scattered points as small rectangles (red/blue). Three axes (X,Y1,Y2) can be displayed in a normal
plot.
Data Selection/Manipulation on a cross-plot
How To
Data Selection/Manipulation Crossplot Window
There are several data selection and manipulation tools available that cover
the following:
Interactive display of selected points in a live scene
Extraction of body or picksets data from selected points
Removal of unwanted points from the cross-plot so that an appropriate data
statistics/relationship is predicted (e.g. regression analysis)
Interactive identification of selected points in a spreadsheet so that the points can be
edited in the sheet.
How To
Data Selection/Manipulation Crossplot Window
Interact mode used for Zoom-In and scroll the original display
selection. If active, then the
remaining selection buttons will
be active.
View mode:
Zoom-In: wheel mouse roll
Pan: left mouse click & drag
How To
Data Selection/Manipulation Crossplot Window
1- Interact mode
4- Removes the selected data (3)
from the cross-plot and the
2- Choose polygon/rectangle
spreadsheet. Use in conjunction
and select points in the
with (1) and (2).
cross-plot (3) by using left
mouse button (drag).
5- Shows selected points in the
spreadsheet so that a user may
edit the corresponding field.
3-Selected data
6- Optional: Selection method can be
logically constrained by using any
mathematical expression. For instance, if
you want to select only the values that are
above 25000, you can write an expression
as x0 > 25000 where x0 can be any
attribute/log available in the spreadsheet.
How To
Data Selection/Manipulation Crossplot Window
1- Interact mode
Interactive display of
selected data as picks in the
2- Choose polygon/rectangle active scene.
and select points in the
cross-plot (3) by using left
mouse button (drag).
3-Selected data
4- Press this button to show the selected
points in the scene.
5- Press this button+(4) to hide the
displayed points in the scene.
How To
Data Selection Extract Bodies/Pickets
Once the data is selected in a cross-plot and it is displayed
in a scene, it can be stored as a pickset or geo-body.
Right click on any pick to pop-up
the menu.
Store the displayed picks as a pickset/body.
The bodies (red/blue) are extracted from cross-plot The picksets (red/blue) are
extracted from the cross-plot
How To
Data Selection Crossplot Window
Multi-Y selections
Moreover, a user can select multiple
Y-axis within one selected region
(polygon/rectangle)
In case of normal plot, the
highlighted points will be
displayed with a change in color.
Snapshot a plot
How To
Snapshot Crossplot Window
A snapshot tool is available to take an image of
a cross-plot in a file (*.png, *.bmp, *.pdf, *.ps
etc).
Take snapshots of a cross-plot
Outline
Purpose and Benefits
How To ...
Menus/short-cuts
Attributes vs Attributes cross-plotting window
Attributes vs wells cross-plotting window
Attribute data window
Main cross-plot window (settings and manipulation)
Cross-plot Examples
Cross-plot Examples
Extract Picksets
(Supervised Prediction)
Quality Control Tool
Cross-Plots
Property
Predictions/Relationships
Extract Bodies
Cross-plot
A Quality Control Tool
Cross-plots can be used as a quality control (QC)
tool specifically in multi-attribute analysis or in
meta-attribute generation.
By plotting multi-attributes in a cross-plot, a
seismic interpreter gets a feel for relationships
between attributes. This may help him/her to
select useful seismic attributes for further
analysis.
As an example, attributes with small statistical
differences yield similar displays (see next slide).
Cross-plot
A Quality Control Tool: Example-1a
Horizon attribute maps
RMS Amplitude within 80ms window Total Amplitudes within 80ms window
This is a simple example of two amplitude-based statistical
attributes. Both attributes result in similar horizon attribute
maps. Their cross-plot shows a narrow data cloud close to a
straight line revealing a large correlation, hence redundant
information. Including both attributes in a quantitative (multi-
attribute) analysis is not helpful.
Cross-plot
A Quality Control Tool: Example-1b
Instantaneous amplitude vs frequency slope fall Amplitude at 44Hz vs frequency slope fall
The selected region highlights the same object. This is as expected because the selected ranges in both
cross-plot correspond to higher amplitudes over the same frequency slope fall.
Cross-plot
Seismic Prediction
Multi-attributes or attributes vs. logs cross-plots
can be used in seismic predictions. The available
plugins (Neural Network, MPSI, etc) may further
help to enhance seismic predictions. For instance,
Supervised seismic facies prediction
Prediction by using linear regression
Establish a relationship (attribute Vs attribute or attribute vs.
log) to predict a seismic object
Cross-plot
Seismic Prediction: Example-2a
Red picksets
This is an example cross-plot of a prominent
decomposition 44Hz
SWFFT Spectral
bright spot. The data is spectrally decomposed
in to discrete frequencies around the bright
spot and the iso-frequency attributes are cross-
plotted.
Blue picksets
SWFFT Spectral
decomposition 24Hz
In this cross-plot note that there is a clear separation
between the two spectral decomposition results at higher
amplitudes (above 20000). Each separation highlights
different parts on a horizon map (Red & Blue picksets)
Interpretation: both parts represent gas-fill. The
separation is a result of tuning effects related to the
thickness of the gas-column.
Horizon attribute map with extracted picksets (red & blue)
Cross-plot
Seismic Prediction: Example-2b
Use regression analysis to predict the seismic-rock
properties. For instance, you can apply a conventional
porosity prediction method (a regression equation).
1- Cross-plot logs (PHI vs Impedance)
2- Create a user defined regression equation
3- Use Mathematics attribute to predict porosity from
available seismic impedance volume.
Seismic Impedance Porosity
10 18 *1000 gm/cc * m/sec 0.1 0.4 gm/cc
Cross-plot
Object Detection
The data selection option can also be used to
extract 3D-bodies and picksets.
Subsequently the extracted bodies and picksets
can be used for e.g.
Reservoir characterization
Prediction
Visualization/Interpretation
etc
Cross-plot
Object Detection Example-3a
Several 3D-bodies are extracted
from cross-plot tool.
Extracted geo-bodies
(Gas pockets red)
Right-click pop-up menu Extracted salt
geo-body
Cross-plot
Object Detection Example-3b
A simple object detection example by using cross-plot tool and Neural Network Plugin
Input horizon map Output facies map
Supervised Neural Network
Multiple selection
(CTRL+Mouse drag)
Class-1 High amplitudes (from cross plot)
Class-2 Moderate amplitude (from cross plot)
Class-3 Weak Amplitude (from corss plot)
Class-4 Faults/edges (manual picks from coherency/dip attributes)
Thank You
