Model Targets - Batch Mode

Use the Batch Mode option (geosoft.uxo.gxnet.dll(Geosoft.GX.UXO.UxaBatchFit;Run)) to invert data from standard or mapping grade sensors (i.e., Magnetics and EM61) for a series or batch of targets and save the results in the target database.

Modelling Parameters - Dynamic Survey Mode dialog options

Main Information
Sensor type	Select the sensor type from the list of available types: EM61, EM61MK2, EM63, EM61MK2-Array, Mag Script Parameter: UXANALYZE.SENSORNAME
Database	Select the method of collecting site data for standard sensors. Select Chronological to view the line profile of your targets. Select Non chronological if no line profile is required. Script Parameter: UXANALYZE.CHRONOLOGICAL [1 - chronological; 0 - non-chronological]
Subset database	Select whether to update or use the existing subset database. The Update option extracts data from the site database and saves the data to the subset database. The Use existing option uses the data directly from the subset database if it exists. If it does not exist, it will extract the data from the site database. Script Parameter: UXANALYZE.UPDATESUBSETDB [1 - update; 0 - use existing]
Site Database
Name	Specify the survey database name. Script Parameter: UXANALYZE.GDB
Sensor channel	Select the sensor channel from the list of channels in the servey database. Script Parameter: UXANALYZE.SENCHAN
Sensor height - Variable channel	Select the sensor height channel from the list of channels in the site database. Script Parameter: UXANALYZE.ALTCHAN
Sensor height - Constant	Specify a constant height applied to all the site data. Script Parameter: UXANALYZE.CONSTHEIGHT
Pitch channel	Select the pitch channel from the list of channels in the site database. This entry is only visible when the sensor type is EM. Script Parameter: UXANALYZE.PITCH
Roll channel	Select the roll channel from the list of channels in the site database. This entry is only visible when the sensor type is EM. Script Parameter: UXANALYZE.ROLL
Heading channel	Select the heading (yaw) channel from the list of channels in the site database. This entry is only visible when the sensor type is EM. Script Parameter: UXANALYZE.HEADING
Coil channel	Select the coil channel from the list of channels in the site database. This entry is only visible when the sensor type is EM61MK2-Array. Script Parameter: UXANALYZE.COILCHAN
Target Database
Name	Specify the target database name. Script Parameter: UXANALYZE.TARGETGDB
Group	Select the target group from the list of groups in the target database to invert. Script Parameter: UXANALYZE.TARGETGROUP
ID channel	Select the ID channel from the list of channels in the target database. Script Parameter: UXANALYZE.TARGETID
Start ID	Specify the starting ID value. Only the targets with the ID equal to or greater than this value will be processed. Script Parameter: UXANALYZE.STARTID
End ID	Specify the ending ID value. Only the targets with the ID equal to or less than this value will be processed. Script Parameter: UXANALYZE.ENDID
Mask channel	Select the target database mask channel to filter out specific targets. Only the targets with a mask (channel value) of 1 will be processed. If the value is set to 0 or dummy (*), the corresponding target will be ignored. If the field is left blank, no mask will be applied. Script Parameter: UXANALYZE.TARGETMASK
Sensor Parameters Information; Mag
Inclination	Specify the inclination value. Magnetic parameters can be calculated via the IGRF GX. Script Parameter: UXANALYZE.INCLINATION
Declination	Specify the declination value. Magnetic parameters can be calculated via the IGRF GX. Script Parameter: UXANALYZE.DECLINATION
Total field	Specify the total field value. Magnetic parameters can be calculated via the IGRF GX. Script Parameter: UXANALYZE.FIELDSTRENGTH
Sensor Parameters Information; EM61
Coil size	Select the coil size from the available options: 1x1 m, 0.5x1 m, or 1x0.5 m. Default selection is 0.5x1 m. Script Parameter: UXANALYZE.COILSIZE
Mode	Select either Mode D or Mode 4. Default selection is D. This entry is only visible for EM61-MK2. Script Parameter: UXANALYZE.MODE
Channels to fit	Select the gate(s) to be used in the modelling process. Script Parameter: UXANALYZE.CHANNELSTOFIT
Coil separation	Specify the coil separation. This entry is only visible for EM61. Script Parameter: UXANALYZE.COILSEP
Coil Geometry
This tab only appears when the sensor type is a EM61 sensor array. The Coil Geometry tab tabulates the coils' locations.
Configuration	The configuration contains the number and location of each sensor in the sensor array, which are read from SensorTableLand.xml. Refer to Application tips below for more information about the Sensor table file.
Set	When selected, the updated configuration is saved into a copy of SensorTableLand.xml located in the %USERPROFILE%\Documents\Geosoft\Desktop Applications \etc folder. In the future execution of this dialog, this new configuration is loaded instead of the original one.
Factory setting	When selected, the original configuration is loaded into the dialog.
Preview	When selected, the sensor array configuration is shown in a schematic map.
[ More ]
Maximum gap in units of average sample spacing	Specify the maximum gap value to select the profile line. Default is 5. This entry is only visible when a chronological database is selected. Script Parameter: UXANALYZE.NGAPS
Maximum deviation angle along the profile	Specify the maximum deviation angle to select the profile line. Default is 90. This entry is only visible when a chronological database is selected. Script Parameter: UXANALYZE.DEVANGLE

Application Notes

For each target in your target database, the magnetic data or EM61 data is extracted and modelled/inverted to determine the source location and parameters. To use the Batch Mode option to invert your data, select the sensor type, and enter the required parameters. At the end of the process the target database is populated with the inverted or fitted parameters.

Your target database should have at a minimum the following channels:

The X and Y location channels that should be of type "double".
An ID channel that should be unique for each target and of type "string", "short", or "long".
A Mask channel with the value of 1 for each target to be modelled/inverted. If you do not want a target to be modelled, set the value to either 0 or dummy (*). This channel should be of type "short".

If an error happens during the batch process, the error messages are logged in batch_fit.log. At the end of the process, you should look at this file for details.

Anomaly Window Polygons

For each target, a polygon window of data about the target location is extracted and modeled. The window should be large enough to encompass a single anomaly along with some background data and should exclude adjacent anomalies. These polygon windows can be defined in two ways:

As a simple square of the size you defined in the 'Anomaly window size' parameter in 'Inversion Settings'.
As a polygon through using the 'Anomaly Polygon' tools – Create, Plot, and Adjust.

Inversion Results

Magnetic Parameters

The inversion of the magnetic data will return the following source parameters in your target database:

Fit_error: Inversion error flag.
     0 - No error
     1 - Anomaly not strong enough to fit
     2 - Not enough data points selected for modelling
    3 - Zero RMS noise
    -3 - Maximum number of iterations reached
    4 - Memory allocation error for target
    5 - Square root of negative number encountered when fitting target
     6 - Division by zero
    7 - Matrix inversion problem
     8 - Parameter values changed too much relative to initial estimates
     9 - Fit process did not converge
Fit_coh: The model coherence, a measure of how well the modelled data fits the measured or observed data; specifically, the correlation coefficient squared (dimensionless). A high fit coherence does not guarantee that the fit results are accurate – only that the match between modelled and measured data is high. Multiple solutions may exist. The fit quality also depends on the number of data samples being passed to the fitting algorithm. The fewer the samples the easier it is to get a high fit coherence.
Fit_ModErr: The model error, a measure of the error in the modelled data (percent). This is calculated by sqrt(1-correlation coefficient squared) *100.
Fit_depth: The depth of the source dipole from the surface (survey distance units)
Fit_x: The X location of the source dipole (survey distance units)
Fit_y: The Y location of the source dipole (survey distance units)
Fit_dec: The declination of the source dipole (degrees)
Fit_inc: The inclination of the source dipole (degrees)
Fit_MagneticMoment: magnetic dipole moment, a measure of the magnetic source (Amp-m2)
Fit_size: Empirically calculated diameter of the object (distance units). The size estimate output from the magnetic data is a function of the object orientation with respect to the earth’s field. The standard deviation of the size estimate is approximately 30% of the ordnance diameter. The apparent size is inferred from the radius of steel sphere which has the same magnetic dipole as the target. In general, the size output from magnetic data is more robust than the size using EM data for common targets. The “fit_size” parameter should be used as a relative indicator of size and not as an absolute indicator of UXO caliber.
Fit_Solid angle: The angle between the earth's magnetic field and the inverted source dipole (degrees).
Index2SubsetGDB: The group number in the subset database holding the corresponding data.
Comments: Initially this string is of zero length - the user may populate it with target related information.
EM61 Parameters
The inversion of the EM61 data will return the following source parameters in your target database:
Fit_error: Inversion error flag.
   0 - No error
   1 - Anomaly not strong enough to fit
   2 - Not enough data points selected for modelling
   3 - Zero RMS noise
  -3 - Reached maximum number of iterations
   4 - Memory allocation error for target
   5 - Square root of negative number encountered when fitting target
   6 - Division by zero
   7 - Matrix inversion problem
   8 - Parameter values changed too much relative to initial estimates
   9 - Fit process did not converge
Fit_coh: The model coherence, a measure of how well the modelled data fits the measured or observed data; specifically, the correlation coefficient squared (dimensionless). Ideally, a fit coefficient of 0.995 or greater is desired for EM data in order to have confidence in the shape information (expressed by the polarizabilities (“betas”) – see definition below. However, if only size and depth are of interest, then a lower fit coherence (i.e. ~0.98) may be acceptable. A high fit coherence does not guarantee that the fit results are accurate – only that the match between modelled and measured data is high. Multiple solutions may exist. The fit quality also depends on the number of data samples being passed to the fitting algorithm. The fewer the samples the easier it is to get a high fit coherence (dimensionless).
Fit_ModErr: The model error, a measure of the error in the modelled data (percent). This is calculated by sqrt(1-correlation coefficient squared) *100.
Fit_chi2: The Chi-Square Statistic is a measure of the error in the model data (dimensionless). It is calculated across the full-time decay for each Transmitter and Receiver combination as ∑((y-yfit)/sdev)².

The value of the Chi Square test-statistic is:

Where:

= Pearson's cumulative test statistic, which asymptotically approaches a distribution.
= observed or measured data
= expected or modelled data
= the number of valid time gates

Fit_depth: The depth of the source from the surface (survey distance units)
Fit_x: The X location of the source (survey distance units)
Fit_y: The Y location of the source (survey distance units)
Fit_theta: Inclination of the source dipole (degrees)
Fit_phi: Declination of the source dipole (degrees)
Fit_psi: Rotation of the source dipole (degrees)
Fit_b1, b2, b3: The three polarizabilities or response coefficients (“betas”), that represent the eigenvalues of the magnetic polarizability tensor. The relative size of the coefficients provides information about the shape of the object. For cylindrical, metallic, and isolated objects, we expect b1>b2≈b3. For spherical objects, b1≈b2≈b3. They can be thought of as components along the three orthogonal axes of the induced dipole.
ModelingChannels: In the case of EM61MKII the channels used to derive the fitted parameters. tip that if multiple channels are modelled, multiple sets of fitted parameters will result. The later source parameters are appended to the bottom of the target group.
Index2SubsetGDB: The group number in the subset database holding the corresponding data.
Comments: Initially, this string is of zero length - the user may populate it with target related information.

Subset Database

For fast access, the process sorts the survey data by X then Y into one big bin. Then for each target, it extracts a window of data centered on the target and of the specified window size. This data is saved in a subset database as:

<Survey database>_<Target database>_<Target group>.gdb

For each target, the extracted data is saved in a random group by Target ID number. In addition to this window, if Chronological is selected, the profile line closest to the target is also extracted. All the targets for which the mask is turned on are then modelled. The calculated model data is saved in the subset database alongside of the observed data, along with the residual – the difference between the observed and model data.

Sensor Lookup Table

The SensorTableLand.xml lookup table file contains various information, which is used to populate the entries in the dialogs and to define the fitting channels. The file is located in the "C:\Program Files\Geosoft\Desktop Applications \etc" folder.

If you would like to customize the file, you should copy the file SensorTableLand.xml to the "%USERPROFILE%\Documents\Geosoft\Desktop Applications \etc" folder; otherwise, your changes may be lost during a software update.

This tool is developed in partnership with Acorn Science and Innovation (AcornSI).

*The GX tool will search in the "gx" folder. The GX.Net tools, however, are embedded in the Geosoft.uxo.gxnet.dll located in the bin folder. If running this GX interactively, bypassing the menu, first change the folder to point to the bin folder, then supply the GX.Net tool in the specified format.