Subtraction of surface-related multiples: adaptive subtraction vs. pattern recognition

Antoine Guitton

antoine@sep.stanford.edu

ABSTRACT

Two multiple subtraction techniques are tested on a 2D synthetic dataset. The first technique is adaptive subtraction, where the signal is assumed to have minimum energy. The second technique is pattern recognition, where the signal and noise are assumed to have different multivariate spectra. Overall, the pattern based technique leads to a better subtraction of the multiples.