Semi-automatic detection of faults in 3D seismic data |
| |
| Authors: | Kristofer M Tingdahl Matthijs de Rooij |
| |
| Institution: | Department of Earth Sciences, Göteborg University, Box 460, 405 30 Göteborg, Sweden;, and dGB Earth Sciences, Nijverheidstraat 11-2, 7511 JM Enschede, The Netherlands |
| |
| Abstract: | The semi‐automated detection of objects has been quite successful in detecting various types of seismic object, such as chimneys. The same technique can be applied successfully to detect faults in 3D seismic data. We show that several different attributes – among others, similarity, frequency and curvature, all of which potentially enhance the visibility of faults – can be combined successfully by an artificial neural network. This results in a fault ‘probability’ cube in which faults are more continuous and noise is suppressed compared with single‐attribute cubes. It is believed that the fault‐cube can be improved further by applying image‐processing techniques to enhance the fault prediction. |
| |
| Keywords: | |
|
|
