The IQR-Based Data Cleaning option has been
transformed into the IQR-based data correction option
The
transformation of the IQR-Based Data Cleaning option, which was implemented in
November 2024, into the IQR-based data correction option, along with improvements
to the correction algorithm, has enhanced the efficiency of using automated
BSEF analysis data to create an attribute section.
As a
result of this transformation, the IQR-Based Data Cleaning settings panel has
been removed from the left tab group, and the settings for interquartile range
data correction have been consolidated into the IQR-based data correction
parameter group. This group is now located on the Primary Attribute panel and
on the settings panel for sections of other attributes:
Thus, for the
primary attribute, which is the real part of the complex relative permittivity,
and for other attributes, it is now possible to individually configure the use
of this option and define correction settings.
When the
checkbox in the IQR-based data correction parameter group is selected, the
georadar profile signals are analyzed during section creation. Based on this
analysis, the section is divided into a specified number of clusters. For each
cluster, attribute range boundaries are calculated using the formula [Q1 - K*
IQR, Q3 + K * IQR], where Q1 is the
first quartile, Q3 is the third quartile, IQR is the interquartile range
calculated as IQR=Q3−Q1, and K is the scaling coefficient for the interquartile
range. The attribute values within each cluster are then recalculated according
to the calculated range boundaries.
The
Q-factor attribute section created without using IQR correction is shown below,
visualization of the section in the overlay mode on the GPR profile. The section consists of two main layers with
significantly different electrophysical properties. Due to these differences,
the boundary between these layers is clearly visible on the GPR profile. The
section reveals high attribute values from the upper layer penetrating into the
lower layer, which is characterized by lower Q-factor values. This phenomenon
occurs due to signal over-reflections from dielectric-contrast objects in the
upper layer into the lower layer.
The section created using IQR-based data correction in the automatic clustering mode shows improved positioning of the section elements relative to the layer boundaries on the GPR profile:
It is
possible to divide the section into clusters not only automatically, based on
the specified number of clusters, but also manually, using a mouse, similar to
the process of creating boundaries on a GPR profile. Below is a video
illustrating user actions during the section creation process using the
IQR-based data correction method in both automatic and manual cluster
definition modes.
If you are viewing this webpage on a home computer or laptop and want to
download the video file, follow these steps: start playing the video, then
right-click within the playback window or on the playback control panel. In the
context menu that appears, select "Save video as...". If you are
viewing the video on a phone or tablet with a touchscreen, pause the video, tap
the context menu icon within the playback window, and select
"Download".