Iterative natural thresholding algorithm for seismic inversion aimed at potentially recovering geological interfaces characterized by weak reflection

Seismic inversion has developed into the core technology of quantitative prediction of rock’s elastic properties of underground in the exploration geophysics. In practical application, seismic impedance inversion is a frequently utilized technique. But it still has unsatisfactory protection of small reflectivity series in real seismic data. These small reflectivity series correspond to the weak reflections of geological interfaces, which is usually a sign of low acoustic impedance contrast in tunnels engineering or reservoirs of oil and gas in petroleum engineering, and other strata of geological significance. Hence, it is very important and meaningful to recover them from original seismic data. This paper adopts a novel algorithm named iterative natural thresholding algorithm (INTA) with the aim to improve the protection of weak reflections. Through an optimal thresholding operator and a concave regularization to enforce an optimal sparse solution, INTA can potentially recover these geological interfaces characterized by weak reflection. We utilize a synthetic seismic data profile and a real seismic data set to test its feasibility and show its better protection of small reflectivity series. From the inversion results, one can see that, compared to common used sparse constrained seismic inversion method, the inverted impedance by the proposed method can better characterize weak reflectors and help understand the structure of the earth more accurately.