A shortest path algorithm for 2D seismic horizon tracking

Seismic horizon mapping is an important step in seismic interpretation. Automatic tracking algorithms are valuable tools since manual mapping can be an extremely tiring and time consuming task. In the present work, we propose a new tracking algorithm for 2D seismic horizons based on shortest paths in Directed Acyclic Graphs. Our approach was designed to balance global and local information of seismic data in order to correctly track horizons at complicated geological scenarios. The main contributions of this work include: a new tracking algorithm for 2D horizons; the definition of a graph dynamically built which models the structure of the seismic horizon; an energy function that balances between local and global properties of the seismic data; and, a new method for building subtraces to improve correlation measurements. The proposed method yielded very satisfactory results when compared to a ground truth horizon provided by an interpreter. Additionally, we compared our results with a similar greedy algorithm, which uses only local information to map horizons. The results illustrated the strength of our design.