Interpreter Driven, Interactive 3D Multi-Attribute Classification
J. Henderson, James A. Lowell, Steve Purves, Gaynor Fisher, David Walraven · 2009
This paper presents a novel 3D seismic attribute classification technique based on a framework in which the interpreter defines the data partitioning required to generate the classified data volume by directly interacting with a multi-attribute visualisation system. The classification process utilises the latest advances in graphics processing unit (GPU) parallel computation technology, enabling the classification to be achieved in real time on standard (non-cluster) desktop workstations. The ability to deliver classification results in real time enables interpreter driven feature definition and introduces a new paradigm in multi-attribute analysis techniques, that can be applied so quickly and easily that they can be incorporated simply into any standard 3D seismic interpretation workflow.