Curvature attribute is useful in delineating faults, fractures, flexures and folds. However, in order to get curvature attributes, it is inevitable to calculate seismic data's second-order derivative which is sensitive to noise and causes unstable and unreliable result. This abstract presents a new 3D curvature attribute analysis method with an excellent anti-noise property. Firstly, continuous wavelet transform is used to extract phase-cosine attribute from seismic data. Then apparent dips are calculated based on complex seismic theory and filtered using median filter. Finally, many curvature attributes with high signal-noise ratio are computed by combining apparent dips in both directions. Synthetic seismogram experiment and real data processing show that the method has two advantages: Firstly, it is insensitive to random noise; Secondly, it can effectively suppress the pseudo boundary information caused by discontinuity between the tilted strata. Therefore, comparing with C3 algorithm, this method is more suitable to data with steep obliquity strata.
Presentation Date: Monday, October 17, 2016
Start Time: 3:45:00 PM
Presentation Type: ORAL