在假设反射系数据服从伯努利-高斯分布条件下,本文运用Bayes方法研究了同时子波估计和反褶积问题,包括地震子波,反射序列和反射序列的统计参数及噪声在内的未知量都被看作是具有适当先验分布的现实化的随机变量,由于确定性方法的计算量相当大,本文采用了一种称为Gibbs采样器的简单蒙特卡罗方法,从未知量的联合后验分布中迭代生成随机样本。为了克服地震反褶积固有的不定性问题,对Gibbs采样器作了一些改进,采用随机样本的简单平均值来近似未知量的最小均方误差(MMSE)估计。同时,为说明本方法的性能,本文给出了几个数值计算例子。 Under the assumption that the reflectivity data obeys Bernoulli-Gaussian distribution, Bayes method is used to study the simultaneous wavelet estimation and deconvolution problems, including the statistical parameters of seismic wavelet, reflection sequence and reflection sequence and the unknown Are considered as realistic random variables with appropriate a priori distribution. Due to the computational complexity of the deterministic method, a simple Monte Carlo method called Gibbs sampler is adopted in this paper, The posterior distribution iterates to generate random samples. In order to overcome the inherent uncertainty of seismic deconvolution, some improvements were made to the Gibbs sampler. A simple average of the random samples was used to approximate the MMSE estimate of the unknowns. Meanwhile, in order to illustrate the performance of this method, several numerical examples are given in this paper.