AI assisted seismic interpretation
Interpreting subsurface faults and salt features in seismic surveys requires many weeks of intensive work from human analysts. Deep-learning models act as an intelligent advisor, rapidly scanning seismic data sets to identify geological features so interpreters can focus their attention on prospective areas. Reducing this time frees the interpreters to focus on areas where their expertise adds greater value. The new systems will accelerate analysis workflows and help to improve production and recovery rates, particularly for complex reservoirs.
The FaultCrawler tool identifies common geological features such as faults. FaultCrawler is now used on seismic datasets across the world and is helping geoscientists to better characterise trapping configurations and fracture networks of potential drilling targets. The SaltCrawler tool uses machine learning to accelerate the velocity model building workflow by automatically interpreting salt boundaries, creating the velocity model and launching seismic migrations. With SaltCrawler, we now have an end-to-end workflow that can dramatically reduce turn-around time for a process that has historically taken 2-3 months.