OpendTect User Documentation version 4.2
Prev		Next

6.3. Cross-plot

The Cross-plot tool is designed to create two dimensional cross plots between 2D/3D seismic data (attributes) and either other attributes or well data. The data can be analysed in multiple maners, using different kinds of colour coding and data selection tools. It may be launched from the menu Analysis --> Cross-plot menu.

6.3.1. Cross-plot data extraction

The crossplot data must first be extracted, either on (a subset of) the seismic grid or along (deviated) well paths. 2D or 3D attributes can be used, and well logs if the extraction is done along the well paths. The extracted data will first be presented in a table that can be saved, before actually selecting the features to cross-plot.

Prompt to choose between 3D and 2D attributes

Prompt to choose between 3D and 2D attributes

The extracted data can be saved in the cross-plot table window (see corresponding section) and reopened without repeating the data extraction, from the menu Analysis --> Cross-plot --> Open.

6.3.1.1. Well based data extraction

This window will extract attributes and/or logs along well(s) paths. The output will be presented in table before being used for cross-plots. At least one well must be selected, and one attribute or one log. It is also possible to select only attributes, or only logs.

The well track and time-depth model provide the locations where to extract the data. Values will be vertically extracted along a specially built measured depth axis. This axis is such that the step between two consecutive depth samples is constant but with few jumps, such that the Z difference (time or depth depending on the survey type) between consecutive depths is around the survey default sampling rate. Therefore at shallow level 1 seismic sample can correspond to 4 meters, then 8 meters at intermediate depths, 12, 16 and so on.

Attribute values are vertically interpolated along that created MD axis, since they are unlikely to be along the Z axis defined by the survey geometry. A polynomial interpolation is performed.
Log values are extracted in the depth domain around the depth to be computed, plus or minus half of the distance to the previous and next depths. All collected values are then processed (upscaled) using a provided "Log resampling method" (see below).

The following specific extraction parameters are available:

Radius around wells: All traces that can be reached within the search radius will be extracted. If several traces around a well are found, the same extracted log value is posted in front of the collected attributes values. This option will only duplicate all data if no attributes are extracted. The default value is the survey bin size, use value "0" to extract only the nearest trace, i.e. one value per well per depth.
Filter positions: See corresponding section
Extract between: The option is used to limit the vertical range of data to extract. The combo boxes list all loaded well markers.
Distance above/below: The option is used to modify the vertical range of the extraction window using a relative distance from the provided well markers, in depth. A negative number will decrease the extraction window, a positive number will increase it.
Log resampling method: Logs will be upscaled using this method. Average should be used for most logs. Median, most frequent and nearest sample are more fit for discrete logs like lithology, but can also occasionally be used for other types.

6.3.1.2. Grid based extraction window

This attribute extraction window will perform a grid based extraction. The output will be presented in a table before being used for cross-plots. At least one attribute must be selected. The list of "Attributes to calculate" represents all attribute currently loaded in the attribute set window, and the stored volume under brackets.

The grid based extraction is performed by selecting location according to one of the following criteria. The minimum cell size as well as the maximum values of all ranges are defined by the survey geometry.

Range: Extracts data on a regular 3D grid. Large default stepouts will decrease the amount of extracted data. Cubes coarser than the requested grid will not be interpolated, undefined values will be posted to the cross-plot table instead.
Polygon: Same as "Range", except that the lateral extend is within an area delimited by a polygon. The lateral cell size can still be defined with inlines and crosslines steps.
Table: Only positions listed in a table will be used for the extraction. The table might be an OpendTect pickset, or a column-sorted ASCII file with inline, crossline and Z values in the first three columns respectively.
Horizon: Extracts data along a 2D/3D horizon, or between horizons. Please note that the attributes will be interpolated if extracted along an horizon, but not if the extraction is done between two horizons (grid based extraction using a user-provided Z step). The "Extra Z" values increase or decrease the extraction window, and work similarly as the attribute set time gates (relatively). The left value applies to the top horizon and the right value applies to the base horizon.

Location filters can be added in order to add one or several restrictions to the area of extraction.

6.3.1.3. Data extraction filters

The filters should be used to further limit the amount of data to be extracted for making cross-plots. Several filters can be used simultaneously.

Filters are of two types: The last four (Range, Polygon, Table and Surface) define extraction areas (see previous section) that will complement the extraction settings provided in the main extraction window. Only points satisfying the main extraction settings and all defined filters will be used for the extraction of attributes.

The first two filters, Random and Subsample, are not position related. The random filter makes a percentage subselection, while the Subselection filter will pass (drop) a user-defined number of samples.

6.3.2. Cross-plot table

This window presents all data collected during the extraction phase. It is primarily used to edit and select the data to be cross-plotted. One line corresponds to one extracted data point, annotated by its position and followed by the collected attributes values (forming at vector, from left to right: logs, attributes, stored data). The star right of the attribute name indicates on which attribute the data is sorted. Empty cells represent attributes than could not be extracted given the settings.

The table enables the manipulation and edition of the collected data, prior to making Cross-plots. For instance it is possible to sort the data from an attribute, and to delete the first or last rows, before sending the data for cross-plot. Please note that the table window is connected to the cross--plot window. Any interactive edition done in the cross-plot window will reflect in the table window that remains open and active while working in the cross-plot window.

[/image]

Standard workflow: Save

, edit, click in a column and then

, click in another column and then

, launch the cross-plot window

.
Note: You can select a column by either clicking on its title cell or by clicking on any single cell.

The tool-bar contains the following edition tools:

Saves the data shown in the table to a file (simple text file or OpendTect object). This format is called position vector data and the data is saved in the survey sub-directory (/Features/*.pvds).

Open/Retrieve the stored data. A short-cut is present in the menu Analysis -> Cross-plot -> Open Crossplot.
Assigns the X-axis of the cross-plot to a selected/highlighted column.
Assigns the Y-axis of the cross-plot to selected/highlighted column.
Removes the selected Y-axis for cross-plotting.
It is used to remove the selected rows in the table.
Moves the selection of the primary Y-axis one column to the left. The cross-plot gets updated accordingly.
Moves the selection of the primary Y-axis one column to the right. The cross-plot gets updated accordingly.
Sorts the selected column to an increasing order, from top to bottom.
Displays selected percentage of the data. If a lot a data was extracted not all will be displayed in the table. Nevertheless the sort and remove lines tools may still apply to all the extracted data, upon request of the user.
Toggles on/off the display of the coordinates.
Toggles on/off the display of Z values.
Shows histogram and statistics on the selected column.
Launches the cross-plot window.

6.3.3. Cross-plot window

The cross-plot window shows the data previously extracted and shown the cross-plot table editor. The window may start empty if insufficient data was select in the table window.

The cross-plot window contain three toolbar to perform various tasks:

Crossplot manipulation toolbar

It is a tool to take the snapshot of the cross-plot.
Gives access to the main crossplot window properties: scaling, statistics, regression line, density plot parameters.
Creates multivariate PDF distributions from the cross-plot data.
Colour code the points with respect to a third attribute value by overlaying a colour.
Colour code the points with respect to the wells they were collected.

Crossplot display toolbar

It is used to update the amount of displayed scattered points in a cross-plot. It increases the efficiency to display the data quickly. Normally, a huge data slows down the machine performance, when displayed as scattered points. To avoid this, this option is used to display a selected percentage of the points. Moreover, the percentage is only set for the data that has values; undefined values are ignored in the percentage.

The scattered points of individual/all wells are displayed.

It is used to show a coloured density cross-plot. Normally, if the displayed scattered data points are too large, it consume huge memory to display the data. In that case, the density plots are useful. In a density plot the data will not be filtered out in the display (% point displayed) as in the case of a scattered point display. In the cross plot properties window, there is a tab available for the density plot. In that tab, a user can set a number (currently displayed data) to display the data. When this plot is ON, there will be a colour table displayed that can be changed. The colour spectrum designates the distribution of the points per pixel.
This option is used to Toggle on/off the second Y-axis (Y2) scattered points. It may be noted that when the second Y-axis (Y+) is selected to be cross plotted against Y1 and X, the data points may become too large to be displayed. Therefore, in this case, the system will prompt a warning to display a given percentage of the data (% points displayed).

Crossplot selection toolbar

It is inactive, if a cross-plot is created with one Y-axis. It allows to make section of the scattered points. The selection settings (Select only Y1/Y2, both) are important to remove unwanted points from the extracted data. When dual Y-axis are cross-plotted, a user can select individual or both Y-axis points by changing this option.
It toggles a pan/selection mode. Selection mode is used to select the scattered points. When pan is toggled on, rest selection options (see below) become inactive, which avoids editing/manipulation of the cross-plotted data.
This option is used to display the selected scattered point in an active scene. The selection of the data points is done by using selection mode. By using this option, the selected scattered data can be saved as a pickset/body. Right-click on an pick in a scene and from the pop-up menu select 'save as a pickset' or 'create body' option.
It is a selection mode (rectangle/polygon). The selection is made by holding the left-mouse button down and selecting the desired area within the cross-plot.
It is used to de-select the selected data points (using selection mode tools).
The unwanted data points can be removed by using selection mode and this trash button. In order to remove the data points of Y1, Y2 or both, use selection mode tools to select an area within which the data is to be removed. After that press this button to remove the data.
It is used to select the data from a cross-plot and display the corresponding selected data in the spread-sheet (Well/Attribute data window).
This option is used to refine the selection by using a mathematical logic over a range. It restricts the selection according to the range set in the Refine Selection window (a pop-up window invokes when this button is pressed). For instance, a user may want to remove a data (x0) within a range of 3-4 from a cross-plot in which 1-2 values are overlapping. To do that, press this button to launch Refine Selection window. In the Enter Ranges field, an equation can be set i.e. X0 > 3 and X0 < 4. Where X0 is desired data in a cross-plot. After this a section can be made within a cross-plot to remove values within the polygon according to the equation.

6.3.3.1. Cross-plot properties

The properties window can be used to adjust the scale, view statistics, add regressing lines etc.

Scaling Tab: Sets the clipping state for each axis, or the amplitude range for the chart. Default: 0, which means that the window is adjusted to fit the entire amplitude distribution.

Statistics Tab: Shows the parameters of a least square fit between the attributes used as X and Y1 (values and errors). The regression line can be displayed in the cross-plot window, as well as the correlation coefficient.

User Defined Tab: Sets a user-defined regression line, that can be displayed as well in the cross-plot window. A simple line can also be drawn in the cross-plot window, in which case the corresponding coefficients will be displayed in this tab.

Density Plot Tab: This tab is used to set the minimum points for the automated density plot. Scatter plots will not be allowed below that number. The tab settings define the bin size prior to counting the number of occurrences.

The following pictures shows an example of density plot. Please note that an additional colorbar has appeared. Units are the number of points that correspond to the colour.

6.3.3.2. Probability Density Functions (PDFs)

Probability Density Functions (PDFs) can be created from the cross-plot toolbar icon

. This icon launches a pop-up dialog that can be used for selecting attributes in order to create PDFs. The PDFs are stored in OpendTect Format, that can later be used for running Bayesian classifications.

Please note that all attributes from the table can be selected. Attribute ranges are generated automatically to fit the extracted data distribution. These can be edited before creating the PDF.

6.3.3.3. Overlay from a third attribute

Scattered points can be coded with respect to the amplitudes of an attribute using that option. The popup window requires the selection of that third attribute, and colorbar specifications (type and amplitude range)

6.3.3.4. Well-based colour-coding

The scattered points in the cross-plot window can be coded with respect to the wells along which the data points were originally extracted. The following utility window can be used to control the colour associated to each well:

The configuration above gives the following result:

6.3.3.5. Selection settings

The selections made interactively in the cross-plot window can be further refined and managed in this window.

The selected scattered points can then be displayed in the active scene by clicking on OK. This allows an interactive display of the cross-plot in a scene. The displayed points (i.e. picks) can be saved in the OpendTect survey either as a Pickset or as a Body. Right-click on any point in the display, it will launch a pop-up menu (see below).

Properties: It can be used to edit the size of the points.
Create Body: Create a Body from the selected picks.
Save as Picksets: To store the selected points as a pickset.
Remove selected points: Removes the selected data.

Prev	Home	Next
Volume Builder Setup		Tie Well to Seismic