Geoscience Reference
In-Depth Information
Multi-trace
Image
Original
Signal
Log power
Spectrum
Filtered
Output
Filter
Supplied
3000
250
1.0
200
0.8
2000
150
0.6
100
0.4
1000
50
0.2
0.0
0
0
0
100
200
300
400
500 -100
0
100
200
0
100 200 300
400
500
600
-300
-200
-100
0
100
200 300
Frequency
Time
Distance
(a) Low pass filter, freq. cut at 10
300
1.0
0.8
200
0.6
100
0.4
0
0.2
0.0
-300
-100
0
100 200 300 400 500 600
Time
-200
-100
0
100
200 300
Frequency
(b) Bandpass, filter, freq. cut at 1, bandpass = 10
Distance
800
600
400
200
1.0
0.8
0.6
0.4
0.2
0.0
-300
0
-200
-400
0
100 200 300 400 500 600
Time
-200
-100
0
100
200 300
Frequency
(c) Bandpass, filter, freq. bandpass = 1-100
Distance
400
1.0
0.8
200
0.6
0
0.4
-200
0.2
0.0
-300
-400
0
100 200 300 400 500 600
-200
-100
0
100
200 300
Time
Frequency
Distance
(d) Bandpass, filter, freq. bandpass = 1-100
fIGURe 7.7
Frequency filtering of a GPR data trace: (a) raw field data, (b) trace with a low-cut de-wow filter
applied, and (c) trace with low-cut de-wow filter and band-pass filter applied.
material is known, then the time-distance GPR cross section can be converted to a depth-distance
cross section, which we can call a pseudo-section. Anomalies on the pseudo-section can be inter-
preted in terms of depth and horizontal location. The ideal situation is for clutter and noise to be
removed from the depth-distance cross section (or three-dimensional display) so that the data are a
true representation of objects in the subsurface. The three steps that can help to achieve this goal are
velocity analysis, migration, and multidimensional frequency-wavenumber filtering. The discussion
in this section is confined to two-dimensional data analysis, but these concepts can be extended to
three-dimensional data sets.
