Seismic DNA – a novel non-local search method for multi-attribute datasets

We have developed a new method for extracting geological features from seismic data that is able to make non-local searches on multi-attribute datasets. The input data, which can be seismic data, seismic attributes, or any other data type, are first transformed into characters. The transformed data are searched for complex features with specific character-signatures using text-based search technology. The method allows for feature patterns of different lengths, with different attribute values, and a variable number of sub-features in the character-signature. To make the design of the character-signatures user-friendly, a visually guided design is implemented that enables the user to compose the character-signatures without knowledge of the underlying technology. Since this search methodology for character-signatures holds many similarities with DNA searches for human genes, it is named ‘Seismic DNA.’