Adaptive Multiple Subtraction Based on Sparse Coding

Jinlin Liu,Wenkai Lu,Yingqiang Zhang

DOI: https://doi.org/10.1109/tgrs.2016.2622399

IF: 8.2

2016-01-01

IEEE Transactions on Geoscience and Remote Sensing

Abstract:Adaptive multiple subtraction is the key step of surface-related multiple elimination methods. The main challenge of this technique resides in removing multiples without distorting primaries. We have developed a new pattern-based method for adaptive multiple subtraction with the consideration that primaries can be better protected if the multiples are differentiated from the primaries. Different from previously proposed methods, our method casts the adaptive multiple subtraction problem as a pattern coding and decoding process. We set out to learn distinguishable patterns from the predicted multiples before estimating the multiples contained in seismic data. Hence, in our method, we first carried out pattern coding of the predicted multiples to learn the special patterns of the multiples within different frequency bands. This coding process aims at exploiting the key patterns contained in the predicted multiples. The learned patterns are then used to decode (extract) the multiples contained in the seismic data, in which process those patterns that are similar to the learned patterns were identified and extracted. Because the learned patterns are obtained from the predicted multiples only, our method avoids the interferences of primaries in nature and shows an impressive capability for removing multiples without distorting the primaries. Our applications on synthetic and real data sets gave some promising results.