【摘要】 通过地震反演数据识别岩性,是地震反演的一项基本任务.由于不同岩性的弹性参数范围常常存在一定程度的重叠,所以给岩性识别带来了很大的困难.本文以叠前反演的弹性参数为基础,通过马尔科夫随机场(Markov Random Field简写为MRF)建立先验模型,按照解释好的测井资料,对不同岩性的弹性参数进行统计,得到计算所需的参数,在贝叶斯(Bayesian)框架下建立岩性分类的目标函数,达到岩性识别的目的.通过马尔科夫随机场建立先验模型,能够建立相邻点间的相互作用关系,得到横向上延续的岩性剖面.本文使用一个楔形模型和Marmousi Ⅱ模型对该方法进行了测试,结果表明,该方法有效可行.同时,本文通过加入误差的方法,检验了反演存在误差对识别结果的影响.更多还原

【Abstract】 Lithologic discrimination by using parameters from seismic inversion is a basic task of seismic inversion.Because different lithologies usually have,to some extent,the similar elastic parameters,it is difficult to identify lithology.To solve this problem,lithologic discrimination method based on Markov random-field is applied.This method firstly builds a priori model through Markov random-field on the basis of elastic parameters of pre-stack inversion,and then obtains Gaussian distribution parameters of iterative computation by means of counting elastic parameters of different lithologies based on interpreted log data and creates objective function of lithologic discrimination under a Bayesian framework,and finally achieves the aim of lithologic discrimination.The priori model can establish interrelationships among adjacent points and obtain continuous lithologic sections.A wedge model and a Marmousi Ⅱ model are used to test the method.Results show that the method is feasible.Meanwhile,the influence of inversion error on lithologic discrimination accuracy is tested by adding error in this paper.更多还原