Concurrent Estimation of Seismic Reflectivity and Q by Using an Optimal Dictionary Learning Method

The seismic reflectivity and quality factor Q play an important role in seismic processing and interpretation, such as improving the resolution of seismic data and enhancing the reservoir identification. Most methods estimate seismic reflectivity and Q separately. However, the error of Q model has a negative impact on the reflectivity estimation and the interference of reflectivity makes Q estimates less reliable. In this paper, we propose a new method for concurrent estimation of seismic reflectivity and Q by using optimal dictionary learning. This new method first constructs a complete dictionary based on the non-stationary convolution model, then computes the reflectivity series under different dictionary matrices with the corresponding referencing Q values, and finally selects the optimal dictionary matrix by comprehensively analyzing the residual and reflectivity sparsity so as to obtain seismic reflectivity and Q simultaneously. The results of synthetic and real data examples test confirm the effectiveness of the proposed method. The proposed method provides accurate estimation of seismic reflectivity and Q, improves the vertical resolution without losing weak events and offers more accurate information concerning stratigraphic features in great details.