Picking of sediment layers using multiresolution analysis

Two fully automatic algorithms are proposed for the picking of sediment layers on subbottom profile. An original approach using a Gaussian pyramid is presented, and compared to a more rigorous one using the wavelet formalism. Both algorithms extract and follow reflectors like the human eye, except in the case of very close reflectors. We then looked for complementary processing, that would enable us to add these kinds of reflectors to processed data. In order to achieve this goal, we study the influence of the wavelet choice on the results, and we are also working on chain selection criteria at high resolution