Directional complex-valued coherence attributes for discontinuous edge detection

We propose directional complex-valued coherence attributes through a simple calculation of the cross-correlation between neighboring complex seismic traces normalized by their corresponding envelope within a local time window along a certain spatial direction. For 3D seismic data with varying directional geological edges, the complex-valued coherence attributes along different spatial directions are distinct, and the coherence along a certain direction can highlight discontinuities at (or near) the perpendicular direction. These separate directional coherence attributes can assist in interpreting the dominant direction(s) of fault development, which is vital in determining sweet spots and locating hydrocarbon wells, and can facilitate the detection of weak or hidden geological edges. In addition, we obtain the minimum complex-valued coherence attribute by comparing all directional coherence volumes to describe the entire lineament and spatial extension direction of geological abnormalities (e.g., channels). In essence, the minimum coherence attribute can be regarded as the result of implementing multi-trace complex-valued coherence calculation along the direction perpendicular to the structural trend. An example of 3D synthetic data with a fault system and channel complex is employed to demonstrate the effectiveness of the directional and minimum complex-valued coherence attributes. The application on a real 3D seismic data of tight sandstone reservoir with faults, flexures and fractures, illustrates that the directional and minimum complex-valued coherence attributes can highlight subtle structures and the directional details of geological abnormalities, which are favorably consistent with the manually interpreted results. •Directional complex-valued coherence attributes are proposed.•The coherence along the direction perpendicular to the structural trend is implemented.•3D synthetic data and 3D real data are adopted to demonstrate the effectiveness of the proposed coherence attributes.•Minimum complex-valued coherence attribute performs better in detecting channel edges, oxbow lakes and small faults.