Total variation regularized poststack seismic impedance inversion based on L1-2 norm

Seismic impedance inversion is a key approach for oil and gas reservoir prediction, and the quality of inversion result directly affects the accuracy of reservoir identification and quantitative interpretation. Given that the stratigraphic impedance usually has blocky characteristics, total variation (TV) regularization with edge-preserving properties has shown excellent performance in seismic impedance inversion. However, the limited ability of TV regularization to sparsely represent the signal and its tendency to produce staircase artifacts during the inversion process, lead to poor accuracy of the impedance inversion result in depicting the geological structure boundaries. In order to enhance the accuracy and reliability of the inversion result, a TV regularization based on L1-2 norm is developed, which provides more sparse inversion result and is less prone to staircase artifacts than conventional TV constraint, when applied to poststack seismic impedance inversion. The objective function of the proposed method can be solved efficiently by difference of convex algorithm and split Bregman iterative algorithm. Both synthetic and real data tests confirm that our method can improve the resolution of the inverted impedance while mitigating the staircase artifacts, thereby realizing the fine depiction of reservoir characteristics, which indicates that it has a broad application prospect.