3D seismic edge detection using magic squares and cubes

Edge detection is a category of geometric seismic attributes that has the capability to delineate vital information from seismic reflection data that can be used to aid qualitative and quantitative interpretation. We have evaluated a new method for geologic interpretation based on templates derived from magic squares and cubes. These are discrete differential operators that approximately calculate the spatial derivative of seismic amplitude through 2D and 3D convolution to locate edges and/or geologic features in seismic data. The new operator benefits from multidirectional scanning leading to efficient detection of different edge locations and their respective orientations. We have tested the new operators against the commonly used Sobel filter using two 3D seismic data volumes. Results of the 5×5×5 magic cube operators provided better definition of seismic features than the 3×3×3 magic cube operators. The overall results compared favorably with the Sobel operator, which suggests that the method can serve as a complementary tool to other existing seismic attributes.