Skip to content

MagNIMBUS Data Correction

VGroshkov edited this page Oct 23, 2026 · 1 revision

Raw total-field data of a drone magnetometer contains, besides the signal of the ground, the magnetic signature of the drone itself: it changes with the heading and attitude of the platform and with the flight altitude, and it differs from line to line. The MagNIMBUS data correction script models these effects and removes them from the TMI series of a MagNIMBUS file.

The script is available for files opened with the MagNIMBUS template.



How it works

  1. Platform model. When Include magnetic is on and the file has the Bx, By, Bz vector channels, the script fits a Tolles-Lawson-type model of the drone's magnetic field to the whole record, using the direction of the vector field and, when Include heading is on, the Heading column. The modelled platform effect is subtracted from TMI. If the estimated effect varies by more than 20 nT (standard deviation), it is considered unreliable and is not applied (see Messages).
  2. Line model. For each survey line, defined by the Next WP column, the script fits a second-order polynomial model of TMI on the selected inputs: position, altitude above ground, altitude above sea level and accelerometer. The model describes slow changes along the line that do not come from the ground. The 10 % lowest and highest TMI values are left out of the fit, so that strong anomalies do not distort the model.
  3. Leveling. The model is subtracted and each line is shifted so that its median equals the mean TMI of the whole file. This also levels the lines against each other.

Lines shorter than 50 measurements, or runs with no line inputs selected, are only leveled (step 3).

The result is written to two new columns:

Column Content
TMI_Fixed Corrected and leveled TMI.
TMI_Fixed_Filtered TMI_Fixed smoothed with a Gaussian filter (σ = 15 measurements).


Parameters

Parameter Default Used in Description
Include position (latitude/longitude) on line model Models changes along the line by position.
Include magnetic (Bx/By/Bz) on platform model Enables the platform model. Needs the Bx, By, Bz columns.
Include heading on platform model Adds the heading to the platform model. Needs the Heading column; has no effect when Include magnetic is off.
Include altitude AGL off line model Models changes by the altitude above ground. Needs the Altitude AGL column.
Include altitude AMSL off line model Models changes by the altitude above sea level. Needs the Altitude column.
Include accelerometer off line model Models changes by the platform acceleration. Needs the AccelX, AccelY, AccelZ columns, which the standard MagNIMBUS template does not have.
Use quantile regression off line model Fits the line model to the median instead of the mean (least squares). Less sensitive to anomalies left in the data, but slower.

Latitude and longitude columns are taken from the template of the opened file automatically. The other columns are read by the names listed above.



Running the script

  1. Open the MagNIMBUS CSV file and select it in the file list.
  2. Crop the takeoff, landing and turns with Select Area → Apply Crop, so that only the survey lines remain. Takeoff, landing and turns contain strong platform effects that the model is not meant to describe, and leaving them in usually makes the platform model exceed the 20 nT limit.
  3. In the "Scripts" tab pick MagNIMBUS data correction, check the inputs to include and click "Apply".
  4. Select the TMI_Fixed or TMI_Fixed_Filtered series and compare it with TMI on the chart and in the grid.

The new columns are not saved to disk until you save the file, and a run can be reverted with Undo. Running the script again replaces both columns.



Messages

Message Meaning
"Platform fit: estimate ... nT is above the 20 nT limit and was discarded. Trim takeoff, landing and turns, then run again." The platform model was not applied. Crop the record to the survey lines and run the script again.
"Platform fit: no vector channels, skipped" The file has no Bx, By, Bz columns; only the line model is applied.
"Platform fit: record too short, skipped" Fewer than 50 TMI values; only the line model is applied.
"Line range is too small" A line is shorter than 50 measurements and is only leveled.
"Warning: No model components selected" No line model inputs are checked; lines are only leveled.
"Percent of dataset used for model: ..." Share of the line used for the fit. Values well below 50 % mean the line is dominated by anomalies.

Clone this wiki locally