A Simple Algorithm for Seismic Skeletonization

Seismic skeletonization is a pattern recognition method that is used to automatically track seismic events and produce an automatic line drawing. By using this method, the cycles in every seismic trace are first detected and the attributes of cycles are characterized. Then the seismic events are tracked based on the minimum of cost function of cycle correlation in three neighbor traces, and the attributes of events are characterized, and the attributes are kept track of in a catalog, according to which, one can perform attribute-based analysis and data filtering and quantitatively interpretation. In this paper, the previous seismic skeletonization is modified and simplified and an quick automatic tracking of events and line drawing can be realized. The experiments on the synthetic data and the raw data show that the modified method is more efficient facilitates the calculation.