Adaptive multiple subtraction using regularized nonstationary regression

Spitz test

simon
Figure 14. Benchmark test on multiple attenuation from Spitz (2007). a - Input synthetic data containing a signal (primary reflection) and noise (multiple reflection) events. b - Model of the noise containing phase differences with respect to the corresponding part of the data. c - Extracted signal. d - Estimated noise.

The next benchmark example (Figure 14) reproduces a test created by Spitz (2007), which simulates a seismic gather after migration or normal moveout correction. The signal (primary reflection) event is horizontal while the noise (multiple reflection) event exhibits curvature and overlaps with the signal at small offsets (Figure 14a). Multiple prediction (Figure 14b) contains a curved event but with incorrect curvature. As in the previous examples, non-stationary regularized regression correctly predicts the noise signal (Figure 14d) using match filtering and produces an accurate signal estimate (Figure 14c).

Adaptive multiple subtraction using regularized nonstationary regression