↑ Top
Table of Contents

Creating grids (G07)

Creating a regular grid from a set of line or point data is a fundamental operation with geophysical data. The INTREPID Gridding tool is a powerful means for creating grids. This guided tour conducts you through a simple gridding operation, showing you how to create a grid from a set of line data.

The following illustration shows the final result of a gridding process in the Gridding tool window.

gtcjgrid01.png

Overview

The Gridding tool enables you to produce a grid (raster) dataset from a field of a line or point (vector) dataset.

You can use the Gridding tool to produce grids from:

  • Point data (eg., gravity),
  • Line data (eg., aeromagnetics)

including both:

  • Regularly spaced data (eg., aeromagnetics) and
  • Variable density sampled data (eg., ground and marine gravity).

The INTREPID Gridding tool can

  • Create grids from point or line data using Minimum Curvature refinement
  • Create grids from line data using Bicubic Splining interpolation
  • Create grids from point or line Falcon and FTG gradient data using Spherical interpolation (SLERP) followed by MITRE refinement.
  • Process very large datasets, using tiling if required
  • Create multiband grids from several line or point dataset fields.
  • Create a single grid from multiple input sources
  • Create a single grid from many datasets ( over 7000 datasets have been used to create one grid).

Gridded data form the basis for image-based enhancements such as:

  • Spectral domain (FFT) and convolution filtering,
  • Interpretation tools (Euler deconvolution, grid-based depth methods),
  • Image-based hard copy products.

Location of sample data for Guided Tours

We provide two complete sets of sample datasets, one in INTREPID format and one in Geosoft format. INTREPID works equally well with both formats. When you want to open a dataset, navigate to the directory containing the required data format.

Where install_path is the path of your INTREPID installation, the project directories for the Guided Tours sample data are
install_path/sample_data/guided_tours/Intrepid_datasets and
install_path/sample_data/guided_tours/Geosoft_datasets.

For example, if INTREPID is installed in
C:/Intrepid/Intrepid 6.0.4.304babe910c_x64
then you can find the INTREPID format sample data at
C:/Intrepid/Intrepid 6.0.4.304babe910c_x64/sample_data/guided_tours/intrepid_datasets

For more information about installing sample data, see “Sample datasets – installing, locating, naming” in INTREPID Guided Tours introduction (G01)

For a more detailed description of INTREPID datasets, see Introduction to the INTREPID database (G20). For even more detail, see INTREPID database, file and data structures (R05).

Location of sample data for Cookbooks

In the adjacent folder to Guided Tours data is a collection of more exotic geophysics datasets and grids already prepared for the cookbook training sessions. You may also gain some insights into the capabilities of the software by testing the Project Manager’s ability to preview and describe the attributes of each of the cookbook datasets.

For an introduction to the Cookbooks, see Using INTREPID Cookbooks (C12).

For a list of available Cookbooks, see Cookbooks.

Sample data for batch processing

Yet more sample datasets are available in install_path/sample_data/examples/datasets. We have provided these for users to learn about batch processing with .task files. For more information, see:

Context of this guided tour

In the context of your data processing cycle, gridding can be used at any stage to visualise data. It is particularly useful as a rapid means of assessing data quality.

Should you complete this guided tour?

This guided tour is intended for introductory level users and contains full detailed instructions. The gridding process it demonstrates is a fundamental INTREPID process. You should complete this tour as part of a thorough evaluation of INTREPID. In later guided tours you can use the Gridding tool to produce grids from the results of the exercises you complete.

What you will do

Flow Chart Summary

Steps to follow

In these steps we refer to the folder Albury. This is a subfolder of {install_path}/sample_data/guided_tours/intrepid_datasets

  1. Launch the Gridding Tool.
  2. Start Project Manager.

    Navigate to the directory {install_path}/sample_data/guided_tours/intrepid_datasets/Albury.

    From the Gridding menu, launch Gridding.

    gtcjgrid02.pnggtcjgrid02a.png

    The INTREPID Gridding tool window appears, showing the Input tab. The Input tab is one of four, located near the top of the Gridding tool window.

    gtcjgrid03.png
  3. Open the input dataset and field
  4. Choose Add. The Open Input Dataset dialog box appears. Select the dataset albury..DIR and choose Open. INTREPID opens the dataset and displays a coloured thumb sketch of the line data in the Input Vector Data panel.

    gtcjgrid04.png

    In the central panel, INTREPID displays the X and Y field alias names and their associated Datum and Projection.

    Below Y Field is the Data field. This is the field that we use to create the grid. From the Data dropdown list, select the field rawmag.

    gtcjgrid05.png
  5. Identify the traverse lines and the tie lines
  6. When you create a grid from geophysical line data, it is normal to include the traverse lines and exclude the tie lines from the grid. We do this because there can be residual location errors that might show up at the crossings of the lines, such as small spikes. It’s better to avoid this. There are two ways of identifying traverse and tie lines.

    • By LineType. This is a special INTREPID field which is set to 2 for traverse lines and 4 for tie lines.
    • By line bearing. This is automatically calculated by INTREPID.

    Whether you use LineType or line bearing to identify traverse and tie lines depends on what type of gridding algorithm you use:

    • If you are using Nearest Neighbours filling and Minimum Curvature refinement you may use either LineType or line bearing as the line identifier.
    • If you are using Bicubic Splining interpolation you must use line bearing as the line identifier.

    Note: The LineType field must have been created before you use the Gridding tool, but the Gridding tool is able to calculate the line bearing automatically.

    For the Albury data you are using, there is already a LineType field in the dataset, so you do not need to create one.

    The default choice is Acquisition Lines identified by LineType. The dropdown list should already show linetype so you do not need to change this setting.

    gtcjgrid06.png
  7. Select the gridding method
  8. Select the Gridding Method tab. The Gridding Method dropdown list shows the different gridding methods available. For this exercise use Nearest Neighbours, which is a method of transferring data values from the geophysical line data into the empty grid.

    The Extrapolation Limit controls the amount of data extrapolation across data gaps and at the grid edges. Leave the default setting as 5.

    gtcjgrid07.png

    Note: For more information about the gridding methods we provide, see “Gridding Method” in Gridding (T22a).

    Note: Operations such as Minimum Curvature are for smoothing the gridding results, not for gridding. See “Minimum Curvature (MinQ) and enabling SLERP” in Gridding (T22a)

  9. Select the grid refinement parameters
  10. Go to the Grid Refinement tab.

    gtcjgrid075.png

    For this exercise, use Minimum Curvature grid refinement. You do not need to change any of the default parameters in this tab.

    Minimum Curvature (MinQ) refinement is a smoothing process that INTREPID repeats according to the number of iterations you specify. Each iteration produces a change in some cell values. The changes become smaller with each iteration. When the maximum change for any cell falls below the Maximum Residual or INTREPID has completed the required number of iterations, the Minimum Curvature process stops.

    For more details, including an explanation of the other parameters, see “Minimum Curvature (MinQ) and enabling SLERP” in Gridding (T22a).

    Note: For tensor data, INTREPID applies MITRE smoothing. This uses the 3D gradient physics of tensors to minimize parts of the observed tensor gradients that are not consistent. For more information, see “MITRE” in Tensor and gradient grid notes (R35).

  11. Select the output grid parameters
  12. Select the Output Grid tab. We want to change the output grid name. To do this choose the Browse button on the far right of the Output File Options. The Output Grid dialog box appears.

    gtcjgrid08.png

    Choose [...] and navigate to the nearby directory Albury. Specify the file name albury_rawmag1 in the text box, choose Save. The Output Grid name updates to the new name.

    Next we want to set the grid cell size. This is really the most important parameter, since it controls the resolution and the size of the output grid. In the Grid Dimensions panel, under Cell size, use the up and down arrows to set the cell size. A good rule of thumb is to set the grid cell size to be 1/4 of the acquisition line spacing for scalar field measurements without horizontal gradients.

    This tool has several enhanced gridding methods that make use of horizontal gradients, if you have them, so that the cell size can be taken up to 1/10 the line spacing. To do this systematically, the sign conventions for the observed gradients must be carefully respected.

    If you don’t know the acquisition line spacing you can measure it using the Survey Path editor or the Visualisation tool. The guided tours introducing these tools each include a line spacing measurement activity (See Visualisation tools (G05). The line spacing for the albury dataset is approximately 200m, so 50m cell size is a safe choice. Enter 50 in the X cell size box, and also press return, to force this to be noticed by the Y cell size box as well. Rectangular rather than square grids are supported, but rarely used.

  13. Proceed with the gridding.
  14. Choose Apply in the bottom right hand corner of the Gridding tool window. INTREPID starts gridding the data. A progress popup image shows you the state of the gridding process.

    gtcjgrid09.png

    When the process is finished, an Information box appears telling you that the data has been successfully gridded. Choose OK.

    gtcjgrid10.png
  15. Exit from the tool.

To exit from the Gridding tool, choose Exit from the File menu.

For a better view of your grid

You can use 3D Explore to examine your grid. Instructions for this can be found in Getting Started with 3D Explore (G23).

Key points for this guided tour

In this guided tour you have used the Gridding tool to create a grid from data stored in a line dataset

Frequently Asked Questions

Q : How big can my datasets be?

A : The INTREPID Gridding tool supports extremely large grids – the grid size is limited by the hardware and Operating System. The tool offers a tiling system that allows computers of modest size to grid very large datasets.

Q : Can I import or export my favourite grid format?

A : INTREPID IO API grid formats are already compatible with ERMapper,ARC/INFO raster, ZYCOR zmap, GMT, Geosoft formats. The INTREPID Export tool can output your data in a wide range of formats including ASCII, Geosoft, ECS, Geosolutions, GEOPAK, NetCDF, and GA (AGSO) formats.

Q : Does INTREPID support multiband grids?

A : Yes, both the INTREPID Gridding tool and other INTREPID grid processing tools support multiband grids.

Q : Can INTREPID deal with ‘holes’ in the data?

A : Yes. Masking is supported. In fact, close management of extrapolation is a critical success factor in many downstream interpretation tools.

Q : What gridding methods does INTREPID use?

A : The INTREPID Gridding tool has four scalar gridding methods (as shown in the table below), a tensor gridding method, and additional grid refinement procedures. After the gridding process (no matter which method you use), INTREPID can refine your grid using LaPlace smoothing and iterative Minimum Curvature refinement.

Method

Purpose

Initial Grid methods

Nearest Neighbour

General purpose – suitable for all data types. The interpolation process honours original data, which makes this method very accurate. Nearest Neighbour is slower than other methods. It is close to a the Watson Delaunay ideas, except that this is not a necessary condition

Bi-Cubic Spline
(Fast Grid)

This method is restricted to line datasets. It uses Bi-Cubic Spline interpolation to estimate grid values. The method is very fast. Extensions for vector components have been made here.

Box Filter

General purpose – suitable for all data types. The Box Filter method is faster but not as accurate as the Nearest Neighbour method.

Variable Density

This method is designed to minimise grid artifacts for datasets which have a variable data separation, eg; land gravity, marine gravity and bathymetry data.

SLERP (Tensor data)

This method starts with a nearest neighbour strategy to identify the 3 nearest observation points to each cell centroid that we wish to estimate a tensor gradient. SLERP, or spherical linear interpolation method is used to estimate the rotational parts of the signal. This is provided for both Falcon and FTG datasets.

Refinement methods

Minimum Curvature

This iterative process smooths the grid points after the initial gridding. It is applied only to interpolated grid values, thus the original data points are honoured.

MITRE

This iterative process will smooth Full Tensor gradient data.

A fifth gridding method, Trend Spine, was available in the past, but it has since been removed. This method used adaptive directions to "follow" the geological strike of features cutting across survey lines.