Seismic Curvature Estimation Based on Combining Gradient Structure Tensor and Multi-window

Geometric attribute, which can be extracted from seismic exploration data, is one of the most important kinds of seismic attributes. Curvature attribute is one of the most useful geometric attribute in 3D seismic data interpretation. It is proved that curvature is related to fault, fracture and oil-gas production. However, curvature attribute is obtained by calculate the partial derivatives of the dip of seismic events. Therefore, estimating dips with high precision is very important. We propose one seismic dip estimating method based on combining gradient structure tensor and multi-window technology, and estimate curvature based on this estimated dips. Firstly, we obtain instantaneous amplitude(IA) and instantaneous phase(IP) through complex trace analysis. Secondly, we construct gradient structural tensor(GST) is based on IP, and do eigendecomposition on GST to estimated seismic dips precisely. Meanwhile, we also utilize multi-window technology to promote estimating precision of seismic dips. Finally, we compute structure curvature (include most-negative curvature, most-positive curvature and so on) based on estimated seismic dips. We verify the effectiveness and precision of our method by apply our method to one synthetic seismic data and two real 3D field data.