Abstract：Starting from Bayes sparse inversion theory, the paper adopted modified Cauchy criterion to build up reflectivity sparse constraint and carry out seismic blind deconvolution processing in combination with math frame of seismic blind deconvolution presented by Mr. Canadas, which can better restore the sparse property of reflectivity sequence, and reach a balance between improving the resolution of seismic data and reducing suppressing weak reflection information at the same time, and using relaxation alternative iteration solving plan to simultaneously obtain reflectivity and mixed phase seismic wavelet. In a situation of available well data or other geologic data, the wave impedance can be introduced as prior constraint of reflectivity, which reduces multi-solution of inverse problem and increases the precision and confidence level of inverse result. The numeric simulation and real cases showed that the algorithm in the paper is characterized by rapid convergence and high precision, which has practical application value.
