↑ Top
Table of Contents

Gridding (T22a)

You can use INTREPID Gridding to convert point and line data to a grid suitable for image processing and contouring, enhancing it using local gradients of the data field. The tool adapts to the datatype of the field you have chosen to grid (Scalar, Vector, Tensor), setting more appropriate defaults as you progress. INTREPID divides the region being processed into a grid of square cells. Each grid cell will contain a value derived from an original data point (an original data cell) or interpolated from neighbouring cells or nearby points (an interpolated cell).

The Pro version of this tool also supports multi-dataset gridding, allowing you to choose many individual survey datasets and dynamically create a grid from the selected channel in each dataset. The current best effort exceeds 7000 datasets.

INTREPID has two other special grid dataset processing tools.

  • The Grid Operations tool can resample the grid, change the cell size and rotate the grid. See Grid Operations (T25) for information about this tool.
  • The INTREPID GridMerge tool can combine overlapping or adjacent grid datasets, correcting for differences between them. See GridMerge—merging multiple grids (T24) for a full description.

In this chapter:

How to use this chapter

Parent topic: Gridding (T22a)

This chapter describes the operation of the Gridding tool. You can use the Gridding both interactively and in INTREPID batch processing mode, using INTREPID task specification files.

Where needed in this chapter, there are separate Interactive and Task files sections. Some sections are marked Interactive only or Task files only.

You can find out how to use the tool and also get background information as follows:

Summary of the gridding process

Parent topic: Gridding (T22a)
  1. If you are gridding line or point data, you specify the data field for gridding in one or more datasets.
  2. If you are enhancing an existing grid, you specify the grid dataset name and go to step 6.

  3. You specify the processes and parameters that you require.
  4. INTREPID performs pre-processing.
  5. INTREPID calculates the number of cells in the grid, based on the extent of the line or point data.
  6. INTREPID calculates centroid values for original data cells.
  7. INTREPID calculates centroid values for interpolated cells (that do not contain original data), using the Nearest Neighbours, Bi-Cubic Spline, Variable data density or Box Filter methods.
  8. INTREPID performs LaPlace smoothing as required.
  9. INTREPID performs the Minimum Curvature smoothing process as required.
  10. INTREPID performs the Masking or Clipping process as required.
  11. INTREPID saves the grid if required, or saves the current tile and goes back to repeat the process for the next selected tile.

Important note about task specification files with this tool

If you are using this tool in batch mode, using a command to execute it with a task specification file, you can use a normal Protobuf .task file.

If you are using this tool interactively, you can only use task specification files in the legacy Parms .job language.

This includes:

  • Loading task specification files for use in an interactive session
  • Saving task specification files to record an interactive session
  • The contents of the HISTORY file

For more information about task specification files, see INTREPID task, HISTORY, report and log files (R06)

You can convert Parms .job files to Protobuf .task files using the Project Manager. See “Converting legacy Parms (.job) files to Protobuf (.task) files” in INTREPID Project Manager (T02)

We are working to address this deficiency. If you would like to discuss it please contact our technical support service.

Using the Gridding tool

Parent topic: Gridding (T22a)

In this section:

Gridding tool—Overview of steps (interactive only)

Parent topic: Using the Gridding tool
To use Gridding with the INTREPID graphic user interface
  1. (If you are launching Gridding from the Project Manager and know which dataset you want to process) Select the dataset for gridding.
  2. Choose Gridding from the Grid menu in the Project Manager, or use the command jgridding.exe. INTREPID displays the Gridding Main window.
  3. gtcjgrid0300314.png
  4. If you have a task specification file, load it using Open Taskfile or Load Options from the File menu. (See Specifying input and output files below for detailed instructions.) If all of the specifications are correct in this file, go to step 10. If you want to modify any settings, carry out the following steps as required.
  5. Choose the input datasets. See Specifying input and output files and Input tab.
  6. Specify the data field for gridding and the geolocation fields. See:
  7. If you are producing the grid from line data, ensure that INTREPID can adequately identify the acquisition lines. See Input tab.
  8. If you are producing a grid from a line or point dataset, specify
  9. Choose Apply. INTREPID performs the gridding process, saves the output dataset and displays the results graphically in the Output Grid panel.
  10. gtcjgrid01.png
  11. If you want to record the specifications for this process in a task specification file so that you can repeat a similar task later, use Save Taskfile or Save Options from the file menu. (See Specifying input and output files for detailed instructions.)
  12. If you want to repeat the process, repeat steps 2–9, varying the parameters and/or data files as required.
  13. To exit Gridding, choose Exit from the File menu.

To view the current set of specifications and a list of the process and files involved, expand the task specification tree in the Process panel on the left.

After Gridding you can carry out a more detailed inspection of the grid using INTREPID Visualisation (see Visualisation (T26)).

You can execute Gridding as a batch task using a task specification file that you have previously prepared. See Displaying options and using task specification files for details.

Gridding scalar data

Parent topic: Using the Gridding tool

If you specify a scalar field as the Data field for gridding (see Input tab—Overview of controls), and you do not specify a gradient field, INTREPID produces a grid of this data.

Gridding vector data

Parent topic: Using the Gridding tool

You can specify a vector field (see “Compound data types” in INTREPID database, file and data structures (R05)) as part of the input for gridding. You can specify it either as the Data (see Input tab—Overview of controls) or as the gradient of the data (see Gridding of potential field data with observed gradients).

jgrouvec.png

You can choose the following derived grid products. The illustration shows the Product drop-down list in the Output grid tab (see Output Grid tab).

The default product is an Enhanced Signal grid. The tool will automatically choose this option, if you use the Advanced button to add an Across gradient field in addition to the scalar signal field you have already chosen.

To produce directly any of the other grid products listed below, you must add, via the Advanced button, the appropriate sets of observed gradient fields.

With the object oriented data types now available in INTREPID, you can also use Single Field to add all vector components via the one field.

The option of having each gradient as a separate field in your database is provided to preserve Canadian thinking, even though users may lose track of what processes have been applied to each component in prior levelling and other data conditioning. .

Product

Details

Analytic signal

“Analytic signal filter (reference)” in INTREPID spectral domain operations reference (R14)

Total horizontal gradient

“Total horizontal derivative filter (reference)” in INTREPID spectral domain operations reference (R14)

East horizontal gradient

Horizontal derivative in the X direction. See “Generalised horizontal derivative filter (reference)” in INTREPID spectral domain operations reference (R14)

North horizontal gradient

Horizontal derivative in the Y direction. See “Generalised horizontal derivative filter (reference)” in INTREPID spectral domain operations reference (R14)

Vertical gradient

“Vertical derivative filter (including fractional vertical derivative) (reference)” in INTREPID spectral domain operations reference (R14)

Enhanced signal

Use this option if you are gridding a scalar field with gradient enhancement. See Gridding of potential field data with observed gradients. This option is not available if you specify only a vector field of components or gradients as the main data for gridding.

Calculate diurnal

This is experimental - beta only

Transverse gradient

Transverse gradient is the Across gradient component

Longitude gradient

Longitudinal gradient is the Along gradient component (nothing to do with geodetic longitudes)

Tilt angle (phase map)

If you select Tilt angle, INTREPID uses the measured gradients to compute

atan(dZ/(total_horizontal_derivative))

It is similar to the analytic signal, except it has the character of the phase of the signal. It varies from –90 to +90 degrees.

See also:

Gridding tensor data

Parent topic: Using the Gridding tool

You can specify a tensor field (see “Compound data types” in INTREPID database, file and data structures (R05)) as the Data field for gridding (see Input tab—Overview of controls).

INTREPID can produce grids of the following products. The illustration shows the Product drop-down list in the Output grid tab (see Output Grid tab).

To access this functionality, you must have created a tensor field in your database, either on import, or via the Spreadsheet tool. Both full tensor and Falcon data are supported.

jgrouten.png

You can choose to produce either a range of one-band scalar grids, or a full tensor grid.

Product

Description

X,Y, Z, XY, YZ, ZX

Individual tensor components

Determinant - I1

First invariant of the tensor

Maximum, Middle, Minimum Eigenvalue

Trace

Txx+Tyy+Tzz

Second invariant - I2

Ratio

see Peterson

Strike

atan(I2/I1)

Horizontal gradient amplitude, direction

Curvature gradient amplitude, direction

Grid

This is the default and it means create a Fully interpolated 6 band BIL Ermapper grid, with each component stacked band by band.

Note, geosoft does not support the notion of multi-band geophysical grids, so do not use a ".grd" output grid extension.

The most important innovation here is the use of the patented SLERP or Spherical Linear interpolation method to allow a Full tensor interpolation. This is used for all of the above options, and the calculation from the interpolated tensor is done at the point of populating the cell with an output value.

Gridding of potential field data with observed gradients

Parent topic: Using the Gridding tool

INTREPID can produce superior results if you supplement a scalar input field with observed gradients.

Specifying gradients as a vector field

If you have a set of gradients stored in a field of vector type (see “Compound data types” in INTREPID database, file and data structures (R05)), follow these steps.

To specify gradients as a vector field
  1. Ensure that you have specified and selected the input dataset (see Specifying input and output files and Specifying several input datasets)
  2. Specify the scalar field that you are gridding. Use the Data drop-down list in the Input tab (see Input tab).
  3. Choose Advanced. INTREPID displays the Input gradients dialog box.
  4. jgrinadv.png

  5. Select Single Field.
  6. Check the Gradients check box and select the gradient vector field from the Gradients drop-down list.
  7. Choose Close.
  8. In the Output grid tab, from the Product drop-down list, select Enhanced Signal.

jgrougre.png

Specifying gradients as scalar components

If you have a set of gradients stored in a field of vector type (see “Compound data types” in INTREPID database, file and data structures (R05)), follow these steps.

To specify gradients as a vector field
  1. Ensure that you have specified and selected the input dataset (see Specifying input and output files and Specifying several input datasets)
  2. Specify the scalar field that you are gridding. Use the Data drop-down list in the Input tab (see Input tab).
  3. Choose Advanced. INTREPID displays the Input gradients dialog box.
  4. jgrinads.png

  5. Select Vector Field Group.
  6. Check the Across, Along, Vertical check boxes according to the gradient fields you have available.
  7. For each gradient field, select a field from the corresponding drop-down list.
  8. Choose Close.
  9. In the Output grid tab, from the Product drop-down list, select Enhanced Signal.

jgrougre.png

The minimum you must specify is one gradient or component of the field. The Bi-Spline method is chosen by default as the appropriate gridding method in this case. It is recommended you not use any Grid refinements such as Laplace or MINQ as these tend to negate the subtle contributions made by the observed gradients.

Enhanced potential field gridding—further notes

Parent topic: Using the Gridding tool

Sign convention for gradients from an aircraft

There is a local coordinate convention associated with acquisition. Since we are dealing with vector components of the potential field, this convention must be applied.

Assuming a moving platform for the acquisition vehicle such as an aeroplane:

  • The local Y positive or length component is tail to tip positive.
  • The cross component is left to right positive.
  • The vertical component is upwards positive.

Storage of gradients in the source dataset

The gridding algorithm accepts any or all of these vector gradients.

Before gridding, convert the differences between readings to nT/m. Divide the difference by the distance between instruments.

The local heading of the observation is also computed from the X, Y data and the components are stored with each scalar field observation as vector components in the E, N & Vertical projection system. That is, the observations are stored in a normalized coordinate frame for internal use.

Opportunities for using this extra gradient data

  • Direct gridding of observed analytical signal
  • Direct gridding of observed total horizontal gradient, i.e.

Akima splines

These are able to take an observed gradient along the spline direction. The line direction gradient component is calculated for each observation point as required.

Minimum Curvature algorithm

Brigg’s formulation is a first order finite difference approximation of La Place condition (13 point kernel). This went to considerable trouble to remove any observed horizontal gradient components.

This traditional formulation is revisited and redone as a second order finite difference approximation (25 point kernel). Provision for observed gradients has also been made.

Variable Density gridding

This alternative method also allows for observed gradients. No implementation supporting observed gradients is available as yet.

The Gridding window

Parent topic: Gridding (T22a)

In this section:

Gridding window diagram

Parent topic: The Gridding window

Here is a diagram of the Gridding window. The table below the diagram has more information about the window elements.

gtcjgrid01dia.png

The following table describes the screen elements.

Element

Purpose

Title bar

Shows the name of the tool.

Menu bar

Enables you to specify input datasets, output datasets, task files and to view on-line help.

Toolbar

Buttons for specifying input and output datasets as well as units of measurement for the gridding parameters.

Task specifications tree

A list of current set of specifications for the process and files involved. To browse the tree use the + and - buttons to expand and contract the sections.

Available filters list

A list of filters available. You can select filters from this list.

Process properties tabs

Properties of the currently selected Gridding process step

Apply button

Click this button to run the filtering process

Panel size adjustment

Use this to change the relative size of the data display area and the lists and properties area

Changing display panel size

Parent topic: The Gridding window

Use the panel size control to change the relative size of the data display area and the task control area.

To adjust the size of the data display and task specification areas
  • Drag the panel size control up or down (or left and right, for specification area).

File menu options

Parent topic: The Gridding window

See Specifying input and output files for more information.

jgrmfile.png

Open Input Dataset Use this option to specify the input vector dataset.

Specify Output Dataset Use this option to specify the grid name. The output grid may be a new or existing grid.

Open Taskfile or Load Options If you want to use an existing task specification file to specify the Gridding process, use this menu option to specify the task specification file required. INTREPID will load the file and use its contents to set all of the parameters for the Gridding process. (See Displaying options and using task specification files for more information).

Save Taskfile or Save Options If you want to save the current Gridding file specifications and parameter settings as an task specification file, use this menu option to specify the filename and save the file. (See Displaying options and using task specification files).

Exit Use this option to exit from the tool.

Toolbar

Parent topic: The Gridding window

The toolbar has the following buttons:

jgrtinou.png
  • Specify input dataset
  • Specify output dataset
  • Open Taskfile or Load Options
  • Save Taskfile or Save Options
  • Degrees display selector
  • Help

Degrees display style

Parent topic: The Gridding window

For latitude and longitude data you can select degrees minutes seconds or decimal degrees using the Degrees display selector button on the toolbar.

Specifying input and output files

Parent topic: Gridding (T22a)

INTREPID has controls for specifying the input and output datasets at logical places in the tool, as well as some controls in the File menu and toolbar.

You can enter the path and ..DIR or .gdb file name of the datasets in the dataset text boxes or browse using the [...] buttons.

If the grid is in the working directory (see “Working directory” in Introduction to INTREPID (R02)), there is no need to enter a path. Paths can be:

  • Relative to the working directory OR
  • Absolute

If you browse, INTREPID displays an Open or Save As dialog box. Use the directory and file selector to locate the file you require. (See “Specifying input and output files” in Introduction to INTREPID (R02) for information about specifying files).

In this section:

Overview of input and output files

Parent topic: Specifying input and output files

This section is a summary of the input and output files, with cross references to the instructions about them.

Aliases

Parent topic: Specifying input and output files

If possible, INTREPID identifies the X, Y and line or point type fields from the dataset aliases. Use the following aliases to identify appropriate fields:

Alias

Field

X

East–West geographic location coordinate

Y

North–South geographic location coordinate

LineType

Line type

PointType

Point quality

See “Vector dataset field aliases” in INTREPID database, file and data structures (R05) for more information about aliases.

Input tab

Parent topic: Gridding (T22a)

The Input tab has controls for specifying input datasets and selecting fields to use in the gridding.

In this section:

See also Specifying input and output files.

Input tab—Overview of controls

Parent topic: Input tab

This section gives an overview of the controls in the Input tab.

jgrinput.PNG

Add, Remove

Use these buttons to add or remove datasets from the Input dataset list. See Specifying input and output files.

Properties

Choose Properties to view information about the selected input dataset (see Specifying several input datasets). INTREPID displays the same information as the Project Manager shows. See “Dataset and file information tabs” in INTREPID Project Manager (T02).

Input dataset

The Input dataset list shows the input datasets currently loaded for gridding. You can select a dataset to view its information and select options for it. See Specifying several input datasets.

X Field, Y Field

These drop-down lists show the selected geolocation fields of the selected input dataset. INTREPID uses the dataset X and Y aliases to make default selections. You can choose different geolocation fields from the lists. See Specifying several input datasets.

Data

This drop-down list shows the data field selected for gridding in the selected input dataset. You can select a different field from the list. See Specifying several input datasets.

Advanced

Use Advanced to specify the gradient fields for gradient-enhanced gridding. See Gridding of potential field data with observed gradients. The selected data type appears immediately to the right of the Advanced button as a hint to the user. The three possible types are illustrated below.

Scalar
jginpscal.PNG
Vector
jginpvect.PNG
Tensor "Coord System"
jginptens.PNG

For multiband datasets the Advanced button changes to a band selector with the total number of bands available shown to the right

Multiband
jginpmbd.PNG

Acquisition lines identified by (check box)

(Line datasets only) Use this check box to specify whether you want to distinguish acquisition lines from other lines in the selected dataset. See Identifying acquisition lines and Specifying several input datasets.

Acquisition lines identified by (drop-down list)

(Line datasets only) Use this drop-down list to select the method of identifying acquisition lines in the selected dataset. See Identifying acquisition lines and Specifying several input datasets.

  • LineType Field (Line datasets only, acquisition lines identified by line type)

When this option is chosen a list box appears to the right allowing the user to choose the appropriate Linetype field name. See Acquisition lines identified by a line type field

  • Nominal bearing (Line datasets only, acquisition lines identifid by nominal bearing)

Nominal bearing for acqusition lines. See Acquisition lines identified by Nominal Bearing and Nominal Bearing.

Heading correction

See Acquisition lines identified by Nominal Bearing and Nominal Bearing.

Specifying several input datasets

Parent topic: Input tab

INTREPID can combine several datasets in one grid. The Input dataset list shows the currently loaded input datasets.

The highlighted dataset is the currently selected dataset. INTREPID displays information about this dataset in the Gridding window Input Tab. You can set input options for this dataset.

To select a different input dataset, click it so that it is highlighted.

Identifying acquisition lines

Parent topic: Input tab

This section is for line dataset gridding only. It is normal practice to use only acquisition lines in the gridding process for line data.

Acquisition lines identified by a line type field

If there is a LineType alias or if you specify a line type field in response to a prompt, INTREPID will use it to automatically exclude non-acquisition line data, processing types 2 and 3 only See “Traverse line numbers and types” in INTREPID database, file and data structures (R05) for a complete list of traverse line types and numbers.

If your dataset does not have a line type field but does have line numbers from which you can derive a line type field, you can use the INTREPID Spreadsheet Editor facility to create it. See “Create new field example—Line Type field” in Spreadsheet Editor (T15)) for details about creating line type fields.

Acquisition lines identified by Nominal Bearing

If you do not specify a line type field, but you do specify a Nominal Bearing (estimate of the strike (orientation) of the acquisition lines), INTREPID will identify the traverse lines oriented within ±22½° of the Nominal Bearing as acquisition lines. INTREPID uses the start and end points of a traverse line to calculate its strike.

If you want to use the Nominal Bearing setting to identify the acquisition lines, but do not know their strike, INTREPID can calculate the average strike of the dataset for you. See Nominal Bearing for instructions.

If you do not identify acquisition lines

If you deselect the tick box next to Acquisition lines identified by, INTREPID will grid all traverse line data.

Point quality

Parent topic: Input tab

This section is for point dataset gridding only. If there is a PointType alias or you have specified a point type field, INTREPID will grid only those points for which the point type = 1. If you have not specified a point type field, INTREPID will grid using all points.

Nominal Bearing

Parent topic: Input tab

This section is for line dataset gridding only. If you are gridding a field from a line dataset using Bi-Cubic Spline, INTREPID needs to know the orientation of the acquisition lines.

If you are gridding a field from a line dataset using Nearest Neighbours, Variable Density or Box Filter, you can specify the orientation of the acquisition lines for the purposes of quality control (i.e., rejecting lines that have different orientation) or to specify rotation for the grid.

INTREPID compares the Nominal Bearing parameter with acquisition line orientation. It uses the parameter for the following purposes:

To calculate the Nominal Bearing

Set the ‘Acquisition lines identified by’ to ‘Nominal Bearing’. Press the Calculate button. INTREPID automatically calculates the bearing of the traverse lines in the dataset for you, and writes the entry into the nominal bearing box.

Subsetting the data

Parent topic: Gridding (T22a)

You can specify a region of interest for the Gridding process using a polygon dataset. INTREPID will only grid data within the region defined.

You can use the Subsection tool to define a polygon (see Subsections of datasets (T21) for full instructions).

To specify a region of interest using a polygon dataset

Select the option Use a subset of the data on the Input tab and specify the polygon dataset required (see Specifying input and output files).

Gridding Method

Parent topic: Gridding (T22a)

Gridding uses a mathematical interpolation and extrapolation process to calculate values for the cells within the edge regions and also within enclosed gaps (fully enclosed by but outside the edge regions). INTREPID ofers a number of methods for this process. You can select and specify parameters for them in the Gridding Method tab. In this section:

Grid initialisation

See Grid initialisation for a description of the initialisation process, common to all gridding methods

General parameters

General parameters are common to all gridding methods. See Gridding method general parameters.

Gridding method—Nearest neighbours

This is our recommended general purpose method. It works well for a wide range of datasets. It uses two-point and three-point planar interpolation (triangulation). See Gridding Method—Nearest Neighbours.

Gridding method—Bi-cubic spline

(Line datasets only) This method uses splining to assign cell centroid values. See Gridding Method—Bi-Cubic Spline.

Gridding method—Box filter

This method is also successful with a wide range of datasets. It does not take account of gradients in the process, and will tend to reduce the magnitude and size of features in your results, compared with the other methods. It uses local averaging to assign cell centroid values (assuming that the original data point is at the centroid of the cell). If you use Nearest Neighbours or one of the Spline methods, INTREPID uses the Box Filter to fill any gaps in the grid after your chosen process is complete. See Gridding Method—Box Filter.

Gridding method—Variable density

This method was designed especially for gravity data acquired at variable sampling densities over the region you wish to create a gridded representation of the field. The basic intent of the method is to estimate various wavelength contents optimally, and to combine them into one, final grid that has the best assemblage of observed wavelengths from the datasets provided.

To illustrate the above discussion further, take the case of regional gravity data being collected on a regular grid at say 4 km spacings. Added to that, you may also have observations taken along roads that criss-cross the area. Added to that again, you may also have some small areas where detailed gravity observations have been carried out on a grid at say 200m spacings.

See Gridding Method—Variable density.

Grid initialisation

Parent topic: Gridding Method

The first stage of gridding involves three processes.

  1. INTREPID reads the input data. If it is a vector dataset, INTREPID creates a grid data structure and aligns it with the source dataset.
  2. INTREPID identifies the grid cells that contain one or more original data points. It finds the nearest data point to the cell centroid and assigns its signal value to the cell. It also records the position of the original data point within the cell for ‘honour original points’ processes.
  3. Note: To be more precise, the coordinates of the sub-cell containing the original data point is recorded for the "honour original points". (There are 128 x 128 sub-cells in a grid cell.)

  4. INTREPID determines the row and column limits for the grid. INTREPID uses the cell extrapolation limit and the original data cells at the ends of each row and column in this process.

Note: The origin of a grid is the centroid of the cell in the top left hand corner.

Gridding method general parameters

Parent topic: Gridding Method

These parameters apply to the gridding process no matter what method you select.

jgrgrmtg.png

Extrapolation limit

Cell Assignment

Save Original Data Points

Save triangles

Gridding method

Gridding parameters summary

The various Gridding processes require different selections of these parameters. The following tables show Gridding parameters that are required for each method.

'Yes' means that you must specify a value.

Blank means that INTREPID does not use or require the parameter in the process at this stage, although its effect on the data may be carried through from a previous stage.

'ON' means that the option must be turned on for the process.

Gridding parameters for line datasets

Process

Grid Cell Size

Extrapolate Cell

Nominal Bearing

Rotate Line Data

Search Radius

Box Iterations

Grid initialisation

yes

yes

if required for acquisition line identification

Nearest Neighbours

yes

yes

Bi-Cubic Spline

yes

yes

yes

ON*

yes

Variable Density

yes

yes

no

yes

Box Filter

yes

yes

if required for acquisition line identification

yes

* for 'oblique' line datasets

Gridding parameters for point datasets

Process

Grid Cell Size

Extrapolate Cell

Box Iterations

Grid initialisation

yes

yes

Nearest Neighbours

yes

yes

Variable Density

yes

yes

yes

Recording the cell centroid and original data point positions

You can save the cell centroids and the 'exact' positions of original data points selected for use (the sample points). We have provided this option for technical auditing of the gridding process.

To save the cell centroid and sample point exact positions

Assign a value to the system parameter INTREPID_HONOUR_ORIGINAL before processing. INTREPID will save a point dataset called honour containing a record of each sample point. Each data point recorded will have two sets of coordinates:

  • The coordinates of the corresponding cell centroid and
  • The 'exact' coordinates of the original data.

If you are examining honour, edit its aliases to select the set of coordinates you require.

See INTREPID system parameters and install.cfg (R07) for further information about system parameters in INTREPID and “Vector dataset field aliases” in INTREPID database, file and data structures (R05) for more information about aliases.

Recording the original data points used

You can save a record of the data points selected for use by this gridding process (the sample points– selected because they were closest to the cell centroids).

To save the sample points used by Bi-Cubic Spline gridding

Assign a value to the system parameter INTREPID_FAST_POINTS before processing. INTREPID will save the sample points for the process as a point dataset called fast.

See INTREPID system parameters and install.cfg (R07) for further information about system parameters in INTREPID.

Extrapolation limit

You must specify cell extrapolation limits for a number of Gridding processes:

  • At the start of the gridding process INTREPID creates edge regions around the original data cells. These regions will contain extrapolated / interpolated data. The edge regions should ideally meet or overlap to satisfactorily fill gaps between original data points (e.g., the gap between traverse lines, or the gap between points of a point dataset). You must specify appropriate cell extrapolation limits for filling these gaps.
  • The image refinement processes also require an edge region containing data around the whole grid. See Grid Refinement tab for details.
  • The data in the edge region around the outside of the grid may be discarded after the image refinement processes. This process is called Clipping.
  • INTREPID interpolates values for all cells within regions fully enclosed by but outside the edge regions. The Masking process deletes (sets to null) these regions if required. Masking needs to know the extent of the edge regions. (See Setting grid nulls for details).
To specify the cell extrapolation limit

Enter the limit (number of cells away from an original data point) into the Extrapolate Cell parameter.

You should normally use an extrapolate cell value of at least 2. The default value is 2.

Cell Assignment

Use this to specify the assignment strategy of observations that fall into the same initial grid cell. This applies only to grids that are created, not reprocessed.

The options are:

  • Nearest Pick the nearest observation to the cell centroid
  • First Use the first observation encountered from the database
  • Last Use the last observation encountered from the database
  • Average Take a running average of all observations in the cell
  • Count Grid up the count of the number of observations in the grid.
  • Minimum Take the smallest observation value as the cell value.
  • Maximum Take the largest observation value as the cell value.

Gridding Method—Nearest Neighbours

Parent topic: Gridding Method

jgrgrmth.png

You can produce grids from line and point dataset fields using this method. It has three stages.

  1. Triangulation:
  2. INTREPID makes a number of passes through the grid. It works along the rows of the grid one by one starting alternately at the most Northerly row working South and the most Southerly row working North.

    For blank cells, INTREPID locates nearby original data cells and uses a triangulation process to calculate values.

    When it locates a neighbouring original data cell, it records this for the blank cell. On the first pass through the grid, INTREPID searches for original data cells that are immediate neighbours. On subsequent passes, INTREPID searches for original data cells one cell further away each time from the blank cell. This process is called shelling. INTREPID searches a minimum of 20 shells around each cell.

    When INTREPID has recorded three neighbouring original data cells for a blank cell, it immediately locates all blank cells whose centroids lie within the triangle formed by the three original data points.

    INTREPID interpolates values for all cells within the triangle and marks all of the cells in the triangle as processed. INTREPID uses an ‘honour original data’ process for this interpolation, since it uses the actual positions of the original data points rather than their cell centroids.

    Note: The honour original data process described here uses the same information but is otherwise not related to the honour original data options provided with minimum curvature image refinement.

    After completing a pass in which it is unable to perform any interpolation, INTREPID moves to stage 2.

  3. Linear interpolation:
  4. INTREPID then passes through the grid examining all of the blank cells for which it has found only two nearby original data points.

    INTREPID interpolates values for all blank cells on the line between the two original data points, once again using the position of the points rather than their cell centroids.

  5. Box filtering:

INTREPID passes through the grid using the Box Filter method (See Gridding Method—Box Filter) and creates values for all remaining blank cells.

Tensor Interpolation using SLERP

A recent invention of Intrepid is to extend the nearest neighbour technique when you are wanting to grid a Full tensor Data field. The Spherical Linear interpolation of the rotations (quaternions) part of the signal is automatically invoked at this stage. The Eigenvalues of the tensor are interpolated using the above Nearest neighbour methods for the current triangle.

In this way, an estimate of the curvature gradients of the Potential field can be made smoothly at any point in the near distance to actual observations.

As this constitutes a major extension to the concept of gradient enhanced gridding, the ability to grid to a finer cell size is also present. Studies show that instead of the normal 4 cells between flight lines, up to 8 cells can be supported without significant aliasing. Testing with the Full Magnetic tensor has not been as extensive as with the Full Gravity gradient tensor.

See “Proprietary tensor processing algorithms (SLERP, MITRE, FTNR and FTGD)” in Tensor and Gradient grid notes (R35)

Gridding Method—Bi-Cubic Spline

Parent topic: Gridding Method

This section is for line dataset gridding only. You can produce grids from line dataset fields using this method, which is sometimes called 'fast grid'.

Note, with lines at an angle it is normal to produce a rotated grid, with either rows or columns in the same orientation as the acquisition lines.

jgrgrmts.png

While using this method of calculating values for interpolated cells, INTREPID creates a located line structure that is closely related to the columns and rows of the grid being created. There are three stages of gridding with this method.

  1. Pass along acquisition lines
  2. INTREPID uses spline curves along the acquisition lines to calculate values that will correspond to all grid cells.

  3. Pass across acquisition lines
  4. In the second pass INTREPID examines each column or row of the located line structure perpendicular to the acquisition line direction. It notes the values obtained in the first pass (nodes) and uses them to calculate a spline curve perpendicular to the acquisition line direction. It uses the spline curve to calculate values in between the traverse lines. The following diagram illustrates the process.

  5. Box filtering

INTREPID passes through the grid using the Box Filter method (See Gridding Method—Box Filter) and creates values for any remaining blank cells.

Bi-Cubic Spline parameters

Minimum, Maximum Scan Distance When interpolating a value for a group of cells during the second pass of the Bi-Cubic Spline process, INTREPID searches in both directions for nodes so that it can calculate a spline curve. It requires two nodes on each side of the cells. This parameter specifies a distance limit for this search. If INTREPID cannot find the nodes within this distance, it will not interpolate the cells.

Specify the Search Radius (in distance units) in the corresponding text box. We recommend a Search Radius larger than twice the line spacing, or 10 times the cell size (The recommended number of cells from one line to the next is 4).

Warning: If your dataset is geodetic (latitude and longitude) you specify the Search Radius in degrees.

Nominal Bearing, Rotate Line Data, Spline Type

Spline Gradient Signal-Noise Blending.

Spline Gradient Noise Level.

Gridding Method—Box Filter

Parent topic: Gridding Method

You can produce grids from line and point dataset fields using this method. The Box filter progressively examines each empty cell in the grid in relation to its immediate neighbours and calculates a value for it based on an average value of its neighbours. The Box Filter process repeats until all cells have a value or until it has completed the maximum number of iterations you have specified.

jgrgrmtb.png

The Box Filter assumes that the original data points are at the cell centroids. It does not take account of the actual position of data points within the cell (i.e., it has no 'honour original points' process) and therefore may not be as accurate as other methods.

Box Filter as a finishing process for other methods

If you use Nearest Neighbours or one of the Spline initial gridding methods, INTREPID uses the Box Filter after the process to fill any remaining gaps in the grid.

Box Filter parameters

Box iterations You can use this text box to specify the maximum number of Box filter iterations you want INTREPID to perform before moving to the image refinement stage.

You do not specify Box Iterations if Box Filter is operating only as a final stage of another initial gridding method.

Gridding Method—Variable density

Parent topic: Gridding Method

jgrgrmtv.png

This method was designed especially for gravity data acquired at variable sampling densities, over the region you wish to create a gridded representation of the field. The basic intent of the method is to estimate various wavelength contents optimally, and to combine them into a final grid that has the best assemblage of observed wavelengths from the datasets provided.

To illustrate the above discussion further, take the case of regional gravity data being collected on a regular grid at say 4 km spacings. Added to that, you may also have observations taken along roads that criss-cross the area. Added to that again, you may also have some small areas where detailed gravity observations have been carried out on a grid at say 200m spacings.

The method is implemented using a multi-grid approach. Working back from the required final cell size, a series of 2 or 3 extra, coarser grids are produced, using the Nearest neighbour technology.

These coarse grids are designed to properly capture the longer wavelength data, at an appropriate cell size. eg for 4 km data, a 1 km cell size would not be aliased, and properly represents this aspect of the signal.

The final cell size might well be 50m, with the aim of capturing the higher frequency, shorter-wavelength field information from the 200m data.

To progress down to this finer cell size, a resampling from the coarsest grid to the next intermediate grid is made, (in the scenario above, to 200m cell size). This intermediate grid is then immediately "weeded" around those cells that fall near an observation of gravity, opening up the gridded representaion of the field for further interpolation and local estimation.

Finally, in the scenario from above, a jump to the 50 m cell size is made.

Notes about implementation detail

  1. Low prime number ratios between cell sizes are used to minimise resampling artefacts. You are asked to choose from a limited set of factors that range through combinations of 2,3,4,5,7 and 11.
  2. Weeding goes out around any observation a distance of 2 cells sizes at each stage.
  3. At least 10 iterations of Minimum Curvature are applied to the coarse grids, to ensure a good representation of the smoothed long-wave content is created and can be handed on to the next stage.
  4. Persistence of longer wavelength features can be clearly seen in the output grids, following application of this method.
  5. The method was devised at Geoscience Australia by several workers, including Alice Murray, in an effort to create a best possible gravity grid from extremely irregularly sampled field data.

Pass list

Reduction factor

This is the list of prime number factors to reduce from a coarse cell size to the final cell size using a multi-grid approach.

Cell Size

Final cell size required

Coarse Iterations

You can influence the number of iterations of coarse MIN. Curvature smoothing.

Save Coarse Grids

We recommend using these for quality control, to confirm you can see the longer wavelengths.

Weight Type

The normal option here is Unity. This means give each point/observation equal weighting during the gridding process. The other options are Magnitude - give more weighting to the observation with the largest positive value eg Depth, and Square Root - give a weighting to the Square Root of the magnitude of the signal during 3 point interpolations.

Grid Refinement tab

Parent topic: Gridding (T22a)

After the initial gridding process, INTREPID can perform LaPlace convolution and Minimum Curvature grid refinement, then restore interpolated grid cells to null as required.

If you are refining an existing grid, INTREPID can perform LaPlace convolution and Minimum Curvature Smoothing refinement. The Minimum Curvature process will not use the honour originals method in this case, since no original data is available.

Specify grid refinement processes in the Grid Refinement tab.

jgrrefin.png

In this section:

LaPlace iterations

Parent topic: Grid Refinement tab

This is a heavy smoothing convolution filter that INTREPID applies to all initial grid estimates. It leaves the original points alone. The process improves the condition of the grid for the Minimum Curvature process. If you perform LaPlace convolution, Minimum Curvature will require fewer iterations to achieve a satisfactory result.

The LaPlace formula appears below.

Where

Z0 is the value in the cell being processed before the adjustment,

Z01 is the value in the cell being processed after the adjustment,

R is the relaxation factor,

Z1, Z2, Z3, Z4 are the values in the surrounding cells as shown,

Z1

Z2

Z0

Z3

Z4

If Z0 does not contain original data, then

Laplace iterations

Contact our technical support service for information about this parameter.

Relaxation factor

See Minimum Curvature for more information about this parameter.

Minimum Curvature

Parent topic: Grid Refinement tab

The Minimum Curvature method progressively examines each interpolated cell in the grid in relation to its immediate neighbours and changes its value based on the value of the neighbours.

The aim of the minimum curvature method is to produce a smooth surface of grid values. Working in both directions, it adjusts cell values using a second derivative calculation based on the differences between the values of the adjacent cells. It does not modify the original data points.

A note about this method and the new Spherical Linear Interpolation method for Full tensor data.

These two methods are largely incompatible with each other. Minimum Curvature is designed to calculate from a scalar measure of a potential field the curvature gradients, in particular Txx,Tyy, Tzz and to use a finite difference method to force the residual of the Trace (Txx+Tyy+Tzz) to zero in all parts of the grid. A Full Tensor grid already contains these curvature gradients and so the above ideas no longer hold. An alternative formulation based on forcing incremental derivatives of the horizontal components into a compliant relationship is in preparation and will be released in Intrepid shortly.

INTREPID uses a relaxation factor to accelerate the interpolation process. As INTREPID calculates a new value for each interpolated cell, it multiplies the difference between the old and adjusted value by the relaxation factor. The default value of the relaxation factor is 1.375.

Where

Z0 is the value in the cell being processed before the Minimum Curvature adjustment,

ZA is the adjusted value in the cell after the Minimum Curvature process has been applied,

R is the relaxation factor,

Z1 is the new value in the cell being processed after the minimum curvature process and relaxation factor adjustment have been applied,

If Z0 does not contain original data, then

INTREPID will repeat the process until it has completed the number of iterations you specify in the Maximum Iterations text box, or the maximum residual change for all cells is less than the value you specified in the Maximum Residual text box, whichever occurs first.

INTREPID normally uses a kernel of cells immediately surrounding the target cell as comparison values in the Minimum Curvature process, as shown below.

The Minimum Curvature process can use an 'honour original data' system if required. If you choose this method, INTREPID allows for the actual position of the original data point in an original data cell (rather than the cell centroid) when calculating the weight (influence) of the cell in an adjustment of a neighbouring cell. If you are using the Gridding tool to enhance an existing grid, the grid will have no link with its original data, and therefore INTREPID cannot use any honour originals method. See below for additional explanation of the honour originals process.

At the end of the Minimum Curvature process, INTREPID can perform several further passes through the grid where it adjusts original data cells values in the same way as it has been adjusting interpolated cells. This process is called smoothing, and has the effect of further smoothing the grid near original data cells.

If you are using the Gridding tool to enhance an existing grid, INTREPID will only perform the Minimum Curvature process in Smoothing mode, since the original data can not be distinguished.

Iterations

Use this spinner to specify the maximum number of times that INTREPID may scan through the grid (number of iterations) during the minimum curvature process.

Kernel size

Use this drop-down list to specify the kernel size for the minimum curvature process.

Maximum Residual

Each time INTREPID scans the grid in its attempt to smooth the grid, it may cause a change in the value of each interpolated cell. As it completes each progressive scan, the change in each cell becomes smaller—the interpolated values are becoming 'settled'.

Use this text box to specify the stage at which you consider that the interpolated values are sufficiently 'settled' (subsequent residual changes will be ignored).

Specify the Maximum Residual in the same units as your signal data.

If you are gridding magnetic data we recommend a value of 0.1nT to produce the best print quality.

The Minimum Curvature process will stop when:

  • Every cell has changed during the last scan by less than the value that you specify (the maximum residual change), or
  • The maximum iterations have been performed.

Relaxation factor

See above in this section for an explanation of the function of the relaxation factor.

(Task files) You can change the value of the relaxation factor if you are using a task specification file for your gridding process. Edit the task specification file, changing the value in the Relaxation_Factor line according to your requirements. Example: Relaxation_Factor : 1.325. See Displaying options and using task specification files for information about batch mode.

Tension

Contact our technical support service for information about this parameter.

Honour Original Data

If you turn off Honour Original Data, INTREPID will assume that the original data points are located at the centroids of all original data cells, and calculate the same weight (influence) for the cell in all directions.

If you turn on Honour Original Data, INTREPID will allow for the actual position of the original data point in an original data cell (rather than the cell centroid) when calculating the weight (influence) of the cell in an adjustment of a neighbouring cell. This can improve the accuracy of the image refinement process. An original data point has one of 128 x 128 different positions in a cell.

If you are using the Gridding tool to enhance an existing grid, original data is not available and INTREPID will automatically turn Honour Original Data off.

The Honour Original Data method has two variants: 1 cell and 2 cell. See the section immediately following for an explanation.

Note: The honour original data process described here uses the same information but is otherwise independent of the honour original data process used in Nearest Neighbours initial gridding (See Gridding Method—Nearest Neighbours).

Honour Original Data—1 or 2 cell

Choose the option button according to your requirements. The default option is 2 cell.

For 1 cell Honour Original Data, INTREPID simply calculates the weight (influence) of each original data cell on neighbouring cells based on the location of the original data point rather than the cell centroid, as explained above.

For 2 cell Honour Original Data, for each original data cell, INTREPID

  1. Locates the quadrant of the original data cell in which the original data point lies;
  2. Treats the 3 cells surrounding this quadrant as original data cells;
  3. Locates the nearest original data points to the respective centroids of the 3 cells;
  4. For each of the 3 cells, uses the value of its chosen data point and its position to calculate the weight (influence) of the cell for adjusting neighbouring cells.

Smoothing

If you turn on Smoothing, after INTREPID has completed the Minimum Curvature process, it will perform several further Minimum Curvature passes through the grid. During these passes it will adjust original data cell values in the same way as it has been adjusting interpolated cell values. This has the effect of further smoothing the grid.

Smoothing Iterations

Use this spin box to specify the number of smoothing iterations to perform.

Save curvature grid

Contact our technical support service for information about this parameter.

Setting grid nulls

Parent topic: Grid Refinement tab

Depending on the density of your original data, the cell size, and the value of the Extrapolate Cell parameter, INTREPID will fill gaps in the grid during the gridding process.

You may, on one hand, require a grid with no undefined regions, and require the Gridding tool to fill all gaps as best it can by interpolation.

On the other hand, you may

  • Have one or more regions which have no observed data at all, and for which interpolated data would be meaningless, or
  • Require that the grid contain only data that is close to original data points.

INTREPID uses a masking process to make the grid conform to these requirements. It sets to null all interpolated cells that are too far from an original data cell.

INTREPID can use the value of the Extrapolate Cell parameter as the limit for retaining or removing interpolated cells. For example, if Extrapolate Cell=2, then it can set to null all interpolated cells that are more than 2 cells away from an original data cell.

Internal—fill holes

If you want INTREPID to fill all holes, check this check box. INTREPID fills all holes regardless of:

  • The value of Extrapolate Cell
  • Your selected Setting grid nulls option for Edges

Edges

Use this drop-down list to specify masking for the edges of the grid.

Setting

Descriptions

Original data limits

INTREPID sets to null all extrapolated edge cells (non-original data cells)

Extrapolation limits

INTREPID sets to null all extrapolated edge cells that are further from an original data cell than the Extrapolate Cell value

Strict Extrapolate cells

A new more robust version of Extrapolation limits which uses a moving window and is sensitive to internal holes; also better for random point data of variable spacing. Three further tuning options are available for this method in batch mode (see below).

Strict Extrapolate Cells tuning options (batch only)

Options available:

  • Minimum Quadrants
  • Only keep an interpolated cell if the minimum number of quadrants were populated with original data (default is 0).
  • Minimum Within
  • Only keep an interpolated cell if the minimum number of original data cells are within the extrapolate limit (default is 1).
  • Quadrant Extension
  • Make quadrants larger so that they overlap (defaults is 0)

Output Grid tab

Parent topic: Gridding (T22a)

The Output tab has controls for specifying the output dataset, its content, datum and projection, rotation, alignment and constraints.

In this section:

Output grid tab—Overview of controls

Parent topic: Output Grid tab

This section gives an overview of the controls in the Output grid tab.

jgroutpt.PNG

Output File Options

Output grid

Use this text box or browse button to enter the name and path of the output grid. See Specifying input and output files.

Bands

Use this spin box to specify the number of bands for the output dataset.

Band

(For datasets with more than one band) Specify the number of the band that you want this gridding operation to create.

Band Name

Set the output grid band name. It defaults to the input Data field name.

Data Options

Null Value

Set the value to be used to represent a null value in the output grid.

DataType

Use this drop-down list to specify the precision of the grid that you are creating. See “Data types in INTREPID datasets” in INTREPID database, file and data structures (R05).

Product

Use this drop-down list to specify the data product that you want to grid.

If you have specified a scalar field with no gradient enhancement, INTREPID disables this list, because it simply grids the data in the data field. See Gridding scalar data.

(If you have specified a vector or tensor field for gridding, or gradient fields (separate components or vector) to enhance a scalar field) Use this drop-down list to select the product for INTREPID to grid. Each type of input data has its own set of entries in this list. For details, see:

Rotation

Use this spin box to specify the number of degrees of rotation for the output grid. See Producing a rotated grid for details.

Grid Dimensions (centroids)

Origin (TL), Bottom Right

Use these spin boxes to specify the X and Y coordinates of the top left (origin) and bottom right points of the output grid. If you set these values, INTREPID automatically adjusts Extents and (number of) Cells.

Note that the coordinates defined are for the centre of the cells at the top left and bottom right of the output grid. They do not define the outside bounding box of the grid ie they are half a cell inside the box limits.

Extents

Use these spin boxes to specify the size in distance units of the grid in the X and Y directions. If you set these values, INTREPID automatically adjusts Bottom Right and (number of) Cells.

Cells

Use these spin boxes to specify the number of cells in the grid in the X and Y directions. If you set these values, INTREPID automatically adjusts Bottom Right and Extents.

Cell size

Use these spin boxes to specify the size in distance units of the cells in the grid in the X and Y directions. If you set these values, INTREPID automatically adjusts Bottom Right, Extents and (number of) Cells. You can specify equal size in both directions using the x=y checkbox. See Specifying the cell size for more information.

If you click and type the required cell size into these boxes rather than use the spinners then you must hit return before INTREPID will update the Bottom Right, Extents and (number of) Cells.

x=y

Use this check box to specify whether cells are of the same size in both directions. If you check this box, INTREPID uses the X Cell size for both cell dimensions.

Reset extents

Choose Reset extents to automatically recalculate the default extents for the output grid.

Producing a rotated grid

Parent topic: Output Grid tab

You may be gridding a field from a line dataset with ‘oblique’ acquisition lines (that do not lie North–South or East–West). The survey may have been created like this, for example, so that the data is perpendicular to the strike of a feature of interest.

You may want to produce a rotated grid from oblique data with its rows or columns in the direction of the acquisition lines. After producing the grid you can rotate it so that the rows lie East–West. See “Rotating a grid” in Grid Operations (T25) for instructions.

If you are gridding an oblique line dataset using the Bi-Cubic Spline or Trend Spline method in the gridding process, you must produce a rotated grid.

With the Nearest Neighbours and Box Filter methods producing a rotated grid from an oblique line dataset is not compulsory. (See Gridding Method.)

Specifying grid alignment

Parent topic: Output Grid tab

This feature is only available via task specification files.

Default

No specific alignment is applied.

Same as another grid

Enter the name and path of the control grid.

Aligned with (another) grid

Enter the name and path of the control grid.

Aligned with coordinates (X, Y)

Specify the coordinates of the alignment point in the X and Y text boxes, using distance units of the output grid.

Specifying the cell size

Parent topic: Output Grid tab

Use the Cell Size spin boxes and the x=y checkbox to specify the size of one grid cell in the distance units of the dataset.

Setting an appropriate cell size is a balance between having a fine 'mesh' grid and having gaps or excessive interpolation between cells which may lead to inaccuracies.

If you are gridding from a line dataset we recommend a cell size not greater than one quarter to one fifth of the line spacing.

If you have a smaller cell size than recommended, we suggest that you increase the cell extrapolation limit (See Extrapolation limit for details).

Warning: If your dataset is geodetic (latitude and longitude) you specify the grid cell size in degrees.

Calculation of the row and column limits

The row and column limits of the grid are lists of the locations of the outermost cells in each row and column which are to contain data.

In this discussion we will use the following notation

EC

= value of the Extrapolate Cell parameter

LO

= cell position of the outermost original data cell at the current end of the current row or column

LE

= cell position EC cells further out than LO for the current end of the current row or column

{LE±EC}

= the set of LE cell positions for the current end of the EC rows or columns before and EC rows/columns after the current row/column

Max{LE±EC}

= the outermost LE position in the LE±EC set

L

= row or column limit finally recorded

INTREPID calculates the row and column limits as follows.

For the left hand end of each row, INTREPID

  1. Locates the leftmost original data cell (LO);
  2. Locates the cell EC cells further left than LO (LE);
  3. Locates the LE positions for the previous EC rows and the next EC rows {LE±EC};
  4. Finds the leftmost cell of the set LE±EC ( Max{LE±EC} )
  5. Records the X coordinate of Max{LE±EC} as the X coordinate of the left hand row limit for the current row (X coordinate of L). (The Y coordinate is, of course, the row number.)

INTREPID performs this process correspondingly for the right end of each row and for each end of the columns.

The overall effect of this process is to

  • Produce a 'square ended' effect for groups of rows and columns, and
  • Fill out rows/columns with 'inadequate' original data cells to a common grid edge.

Apply

Parent topic: Gridding (T22a)

When you choose Apply INTREPID carries out the Gridding process that you have specified.

Displaying options and using task specification files

Parent topic: Gridding (T22a)

This tool has special task specification file requirements. See Using the Gridding tool

Displaying options

To view the current file specifications and parameter settings

In this tool the settings are all visible in the process settings panel.

Using task specification files

You can store sets of file specifications and parameter settings for this tool in task specification files using the current Protobuf (.task) language. For a limited time it will also be possible to use the legacy Parms (.job) format.

The instructions in this section are simple steps for the current tool. For a full explanation and instructions on creating and editing task specification files see INTREPID task, HISTORY, report and log files (R06).

To create a task specification file with this tool

We are currently moving INTREPID task specification files to the new Protobuf system. If the tool’s interactive mode has been migrated to Protobuf, it will create task files in the current .task format. Otherwise it will use legacy Parms .job format. For details about these formats, see “Using task specification files in interactive mode” in INTREPID task, HISTORY, report and log files (R06). For more information please contact our technical support service.

Follow these steps:

  1. Specify all files and parameters.
  2. If possible, execute the task (choose Apply) to ensure that it will work.
  3. Choose Save Taskfile or Save Options from the File menu. Specify a name for the task specification file. INTREPID creates the file with the settings current at the time of saving, and adds the extension .task (or .job) depending on the update status of the tool.

You can convert .job files to .task format using the converter that we supply. For instructions, see “Converting legacy Parms (.job) files to Protobuf (.task) files” in INTREPID Project Manager (T02).

You can run most INTREPID tools in batch mode with Protobuf .task files.

To use a task specification file in an interactive session

We are working to update all INTREPID tools so that they load .task format files for interactive sessions. Some may still only load legacy.job format files. For more information please contact our technical support service.

Load the task specification (.task or .job) file (File > Open Taskfile or Load Options), modify any settings as required, then choose Apply (or, for some tools, Process).

To use a task specification file for a batch mode Gridding task

Type the command jgridding.exe with the switch –batch followed by the name of the task specification file.

For example, if you had a task specification file called surv329.task in the current directory, you would use the command:

jgirdding.exe –batch surv329.task

Settings only available in batch mode

You can modify the following operations or parameters only by editing a task specification file:

  • You can modify the relaxation factor in the minimum curvature process (See Minimum Curvature for details). Change the value in the Relaxation_Factor = line according to your requirements, for example, Relaxation_Factor = 1.325.

Task specification file example

Here is an example of a task specification file for Gridding.

# V6.0 gridding Example task file - # demo of how gridding can take a vector field and also grid to a gravity vector grid # usage jgridding -batch grid_enhanced_gravity.task IntrepidTask { Gridding { Input_Datasets { Input_Dataset: “o_NewType/components..DIR”; Input_X: “east”; Input_Y: “north”; Data: “G_acc”;# this is the vector field, not just a scalar } Gridding_Method { Extrapolation_Limit: 5 Cell_Assignment: Closest Save_Triangles: false use_Barycentric_Interpolation: true# default is inverse distance Name: Neighbours Weight_Type: Unity } Grid_Refinement { Fill_Holes: true Edge_Clipping: Extrapolation_limits Laplace_Iterations: 2 Minimum_Curvature { Iterations: 100 Max_Residual: 0.0 Tension: 0.0 Relaxation_Factor: 1.375 size: K5_5 Honour_Original_Data: true } } Output_Grid { Output_Grid: “./sample_vector_grid.ers” Projection{ map_projection: “NUTM17” spatial_datum: “NAD83” } Cell_Size_X: 100.000000 Cell_Size_Y: 100.000000 Band: 1 Bands: 1 DataType: VectorGrid Rotation: 0.0 VectorProduct: VECTOR_GRID# check out the list of possible products here } }}

Frequently asked questions

Parent topic: Gridding (T22a)

Q : Why doesn't Bispline gridding appear to fill the grid properly on a dataset with angled lines?

Splining only works in the NS or EW direction. It depends on the bearing entered. If you enter a bearing of 90, it will spline (NS) for lines which have a bearing within 70º–110º (±20º from 90º), or if you enter 0 it will spline EW for lines bearing 340º–20º. If you enter 45 you must select Rotate dataset or else it will try and spline NS (as if lines were bearing 90º) and not find many node points, outputting a generally blank grid.

Q : What should I make the coarse grid Maximum scan distance for variable density gridding?

The coarse grid search radius (Max scan distance) should default to something very large in order to effectively scan the entire area.

It is in the same units as the X and Y coordinates.

Q : I want to grid the tie lines as well as the traverse lines. How do I do that?

Use the Nearest Neighbours option if you want to grid all the lines. If you have a Linetype field defined, re-number it to 2 for all lines, then grid.

Another way is to comment out the Linetype alias in the .isi file, then grid. The Nominal Bearing entry will then control what lines get gridded. If you leave it blank all lines get gridded.

Bi-splining won't grid all the lines because it ignores lines > 22º from the required nominal bearing.

Q : How can I change the null value in an INTREPID grid?

Edit the grid using the Spreadsheet Editor, and replace the nulls with the new value.

Then edit the grid header file, and change NullCellValue to the new value.