3D Seismic horizon tracking using an artificial neural network

In the structural analysis of seismic data, event picking and classification are essential, though often difficult. Events on a seismic section are usually recognised as strong reflections of comparable signature, amplitude and two-way traveltime, appearing consistently from trace to trace. An interpreter normally assesses the morphology or character of the reflection, i.e. wavelet shape, and not simply reflection amplitude. An overall view of the regional geology, and knowledge of the stratigraphic and structural relationships of the various geological formations is also essential in the interpretation of horizons, especially in areas that are geologically complex.