AGM 3
2
2
what is the f-k filter generally
a development of the frequency filter
f-k means?
frequency wavenumber
wavenumber?
k, reciprocal of wavelength
wavenumber of fast velocity v.s slow velocity
fast velocity wavenumber close to zero
slow velocity has positive or negative wavenumber
f-k spectrum
.
problem with f-k filter
spatial aliasing
worminess
introduction of cherent events
spatial aliasing is
making one event look like another
worminess is
filtered data looks striped
deconvolution simple explanation
if we know the waveform that was put into the ground, we can filter that waveform and simplify the data
DBS
Deconvolution before stack
what does DBS do
supresses short path multiples
short path multiple
energy which reflects between thin beds
how does predictive deconvoltion work
looks for wavelets that looks like the short path multiple and removes them from the trace
what delay do short path multiples have
a constant one
problem with predictive deconvolution
lag between primary and multiple wavelets can be predicted, if delays between primary arrivals is predictable (eg. sequential depostion) they may end up being deconvoluted.
4 deconvolutions
predictive deconvolution
source signature deconvolution
wavelet shaping
inverse q deconvolution
Typical use of radon filter
filter long path multiple
long path filtering constant delays diagram
radon filering/slant stack filtering
how to suppress long path multiples
predictive deconvolution applied using radon transform.
pre processing algorithms are concerned with
improving the signal to noise ration, in prep for velocity analysis.
4 noise classifications
ramdom or coherant
source generated or ambient
removing taces from seismic data is called
muting
types of muting
refraction mute (at top)/top mute
noise cone mute (cone at bottom)
surgical mute (removal of specific area eg. ground roll)
AGC
automatic gain control
balances amplitudes down a trace (or 'window'-part of a trace)
opposite of the Fourier domain
Amplitude plotted against time
(normal)
Fourier domain
Fourier=Frequency
Amplitude plotted against frequency
why are pre processing algorithms applied?
to prepare data for velocity analysis
supress noise, boost signal
what does 'main processing' acheive?
velocity analysis
what does post processing achieve
migration
(focuses the image)
air blast noise classification
coherent
330m/s
high frequency
weather noise classification
random
ground roll noise classification
coherent
slow
low frequency
cultural noise classification
random or coherent
multiples noise classification
coherent
5 examples of land noise
airblast
ground-roll
weather
cultural
multiples
6 examples of marine noise
multiples
swell
propellor
bubble pules
ocean bottom cables
drilling platforms
different noise classifications
random or coherent
source generated or ambient
type of noise?
bubble pulse
what is amplitude recovery
like adding gain to allow for energy spreading
why is amplitude recovery necessary?
because amplitude decays at 1/r
processes:
geometrical spreading (spherical divergence)
attenuation
-
- 1 / 43
-
Cartes-fiches
long path filtering constant delays diagram
radon filering/slant stack filtering
how to suppress long path multiples
predictive deconvolution applied using radon transform.
pre processing algorithms are concerned with
improving the signal to noise ration, in prep for velocity analysis.
