Artificial neural network classification of texture orientations in seismic images

Texture orientation is a very important attributes used in the interpretation of seismic images. It provides critical clues of continuity and connectivity useful in relating adjacent image areas. The author reports on a novel approach in which stacked seismic data are convolved with directional convolution masks and the results are used as input to an artificial neural network for classification of image areas into a number of discrete texture orientation classes. Test results on a piece of real seismic data from the Gulf of Mexico are shown to illustrate the effectiveness of the approach. The instantaneous responsiveness of a neural net makes this approach very practical in processing a large number of seismic images in which orientation of image events offer important cues needed to link various sections in order to construct a 3D image of the Earth's subsurface.