unary adaptive subtraction of joint multiple models with complex wavelet frames
DESCRIPTION
Multiple attenuation is one of the greatest challenges in seismic processing. Due to the high cross-correlation between primaries and multiples, attenuating the latter without distorting the former is a complicated problem. We propose here a joint multiple model-based adaptive subtraction, using single-sample unary filters' estimation in a complex wavelet transformed domain. The method offers more robustness to incoherent noise through redundant decomposition. It is first tested on synthetic data, then applied on real-field data, with a single-model adaptation and a combination of several multiple models. Read More: http://library.seg.org/doi/abs/10.1190/segam2012-0440.1TRANSCRIPT
Unary adaptive subtraction of joint multiple models with complex wavelet framesTaM0: Non-stationary, wavelet-based, adaptive multiple removalTaM1: “Complex” wavelet transform + simple one-tap (unary) filterTaM2: Redundancy selection: noise robustness and processing speedTaM3: Smooth adaptation to adaptive joint multiple model filtering
S. Ventosa(1), S. Le Roy, I. Huard, A. Pica(2), H. Rabeson, L. Duval*(3)(1)IPGP, (2)CGGVeritas, (3)IFP Energies nouvellesContacts: [email protected],[email protected]
Motivation: Multiple model dataShot number
Tim
e (s
)
1200140016001800200022001.8
2
2.2
2.4
2.6
2.8
3
3.2
3.4
Shot number
Tim
e (s
)
1200140016001800200022001.8
2
2.2
2.4
2.6
2.8
3
3.2
3.4
Shot number
Tim
e (s
)
1200140016001800200022001.8
2
2.2
2.4
2.6
2.8
3
3.2
3.4
Shot number
Tim
e (s
)
1200140016001800200022001.8
2
2.2
2.4
2.6
2.8
3
3.2
3.4
Data and three multiple models, common offset plane (need for a model-based, non-stationary, adaptive multiple filtering).
Complex wavelet frame decomposition•Complex Morlet wavelet definition:
ψ(t) = π−1/4e−iω0te−t2/2, ω0: central frequency (1)
•Discretized time r, octave j, voice v:
ψvr,j[n] =1√
2j+v/Vψ
(nT − r2jb02j+v/V
), b0: sampling at scale zero (2)
•Time-scale analysis:
d = dvr,j =⟨d[n], ψvr,j[n]
⟩=∑
n
d[n]ψvr,j[n] (3)
Time-scale data and model trace representations
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Data and model trace Morlet wavelet scalograms.
Redundancy selection
0 0.5 1 1.5 2 2.5 3 3.5 4Time (s)
primarymultiplenoisesum
0 0.5 1 1.5 2 2.5 3 3.5 4−0.1
−0.05
0
0.05
0.1
0.15
Time (s)
true multipleadapted multiple
Key features• fast off-line parameter selection
– realistic synthetics– varying random noise realizations– SNR-based wavelet parameter selection
• controllable redundancy allows:
– simple stable synthesis dual frame– resistance to field noise– computational efficiency balanced
•Morlet wavelet frame
– approximately analytic– sliding window processing along scales
46
810
1214
16
5
10
15
2010
12
14
16
18
20
RedundancyS/N (dB)
Med
ian
S/N
adap
t (dB
)
10
11
12
13
14
15
16
17
18
19
Complex wavelet domain adaptation
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Time (s)
Sca
le
2.8 3 3.2 3.4 3.6 3.8 4 4.2
2
4
8
16
0
500
1000
1500
2000
Adapted joint and individual model trace Morlet wavelet scalograms.
Unary filter estimation•Windowed adaptation: complex aopt compensates local delay/amplitude mis-
matches:
aopt = argmin{ak}(k∈K)
∥∥∥∥∥d−∑
k
akxk
∥∥∥∥∥
2
(4)
•Vector Wiener equations for complex signals:
〈d,xm〉 =∑
k
ak 〈xk,xm〉 (5)
•Time-scale synthesis:d[n] =
∑
r
∑
j,v
dvr,jψvr,j[n] (6)
Results: field data multiple filtering
Shot number
Tim
e (s
)
1200140016001800200022001.8
2
2.2
2.4
2.6
2.8
3
3.2
3.4
Shot number
Tim
e (s
)
1200140016001800200022001.8
2
2.2
2.4
2.6
2.8
3
3.2
3.4
Subtraction results: (top) model 3 (bottom) joint multi-model multiple removal.
Some multiples better attenuated around 3s, random noises reduced.
References[1] Herrmann, F. J. and D. Wang and D. J. Verschuur, 2008, Adaptive curvelet-domain
primary-multiple separation: Geophysics, 73, A17–A21.
[2] Donno, D., H. Chauris, and M. Noble, 2010, Curvelet-based multiple prediction: Geo-physics, 75, WB255–WB263.
[3] Neelamani, R., A. Baumstein, and W. S. Ross, 2010, Adaptive subtraction usingcomplex-valued curvelet transforms: Geophysics, 75, V51–V60.
[4] Jacques, L., L. Duval, C. Chaux, and G. Peyre, 2011, A panorama on multiscale geo-metric representations, intertwining spatial, directional and frequency selectivity: SignalProcess., 91, 2699–2730.
[5] Ventosa, S., H. Rabeson, P. Ricarte, and L. Duval, 2011, Complex wavelet adaptivemultiple subtraction with unary filters: Proc. EAGE Conf. Tech. Exhib.
[6] Ventosa, S., S. Le Roy, I. Huard, A. Pica, H. Rabeson, P. Ricarte, and L. Duval, 2012,Adaptive multiple subtraction with wavelet-based complex unary Wiener filters: Geo-physics, 77, V183–V192.
SEG Annual Meeting, Las Vegas, Nevada, USA, 4-9 November 2012