Multi-trace basis-pursuit seismic inversion for resolution enhancement

Seismic inversion is an important tool that transfers interface information of seismic data to formation information, which renders the seismic data easily understood by geologists or petroleum engineers. In this study, a novel multi-trace basis-pursuit inversion method based on the Bayesian theory is proposed to enhance the vertical resolution and overcome the lateral instability of inversion results between different traces occasionally seen in the traditional trace-by-trace basis-pursuit inversion method. The Markov process is initially introduced to describe the relationship between adjacent seismic traces and their correlation, which we then close couple in the equation of our new inversion method. A recursive function is further derived to simplify the inversion process by considering the particularity of the coefficient matrix in the multi-trace inversion equation. A series of numerical-analysis and field data examples demonstrates that both the traditional and the new methods for P-wave impedance inversion are helpful in enhancing the resolution of thin beds that are usually difficult to discern from original seismic profiles, thus highlighting the importance of acoustic-impedance inversion for thin bed interpretation. Furthermore, in addition to yielding thin bed inversion results with enhanced lateral continuity and high vertical resolution, our proposed method is robust to noise and cannot be easily contaminated by it, which we verify using both synthetic and field data.