Abstract:Lame parameters are much more sensitive and have an advantage over traditional elastic parameters such as velocity and density in reservoir prediction and hydrocarbon direct indication.Elastic impedance inversion as a practical seismic inversion methodology is extensively used for its amounts of information,high resolution and calculation efficiency.This paper realizes elastic inversion and lame parameter direct extracting with modified Cauchy distribution and spike sparse inversion method based on Gray approximation and Bayes theory.This method is capable to better restore sparseness of parameters and realize the direct abstracting for lame parameters,which leads to less affection of weak signal and less error brought by accumulating calculation compared with conventional elastic impedance inversion methods.Both model experiments and real examples show that this method is stable for research of heterogeneous media such as fracture-cavernous carbonate reservoir.