Pre-stack seismic inversion based on orthogonal matching pursuit algorithm

Abstract

The previous work on pre-stack seismic inversion mostly indicated that there are wavelet overprints in the inverted results. The estimated elastic parameters are changed smoothly in the layers' edge which leads to a weak resolution power of the layers. The goal of this study is to offer a new method of pre-stack seismic inversion to generate the blocky layers with high resolution power. This paper proposes using the joint constraint of the model regularization and L0 norm constraint. Where the model regularization is used to enhance inversion stability and L0 norm constraint ensures the reflection coefficients are sparse. Then this study combines the Zoeppritz approximation derived by Fatti to establish the pre-stack seismic inversion optimization problem of L0 norm. Due to the limitation of matching pursuit this study chooses the orthogonal matching pursuit method to solve the optimization problem. At last, the transformation of the sparse reflection coefficients generated by orthogonal matching pursuit method and the elastic parameters by integration is needed. The elastic parameters will express as blocky or the step stairs because the reflections series is sparse. The method is performed in a synthetic example and a field example. In all of the examples, the proposed orthogonal matching pursuit pre-stack seismic inversion method works well with a credible results compared to the true model and the well log. The results inverted from the field data are blockly which have a higher resolving power to the layers and the blockly consequences are more advantageous in geological interpretation.