Petrospire: Indexing and retrieval of seismic data for oil and gas exploration

Abstract

We describe a novel approach to index a seismic data archive as a database in order to provide sufficient precision and recall in specific tasks such as retrieving seismic data containing faults and strong horizons, which commonly arise in oil and gas exploration scenarios. The results show that 70% recall with 80% precision can be achieved using this methodology. This technology could be deployed as a significant aid to geologists working to interpret or explore seismic data. Seismic data archives present one of the most difficult challenges for search and retrieval from 3D archives due to the requirements of domain specific knowledge and large data volumes for the retrieval tasks.