为解决地球物理反演中多解性的问题,综合多种地球物理信息的联合反演受到了广泛的关注.本文依据不同地球物理响应可能由相同异常体引起,而不同地球物理分布参数之间存在相关性等特点,提出了一种基于局部Pearson相关系数约束的联合反演方法.该方法假设每个局部区域模型参数的分布具有线性相关特性,在拟合不同类型观测数据时,对局部模型参数施加相关性约束,进行联合反演以减少多解性.本文采用交替迭代联合反演流程,改善了同—目标函数下联合反演收敛性和速度问题.基于新的联合反演方法和流程,我们测试了三维大地电磁和重力仿真数据的联合反演.结果表明,本文提出的基于局部相关性约束的联合反演方法,能充分利用大地电磁和重力观测数据信息,有效改善单一地球物理反演收敛性和多解性的问题,反演效果得到明显提升.%Joint inversion of multiple geophysical data has attracted widespread attentions due to its large potential to solve the problem with the non-uniqueness in geophysical inversions.In general,different geophysical anomalies can be caused by a same abnormal body,and the distributions of these parameters are correlated.Thus,we develop a joint inversion algorithm based on local correlation constraints.In this algorithm,which assumes that every local region is linearly correlated,the data of different types are modeled with constraints of the local correlations to realize joint inversion and reduce non-uniqueness.We adopt an alternative joint inversion scheme to improve the efficiency of the inversion.With the algorithm,we test on 3D synthetic magnetotelluric (MT) and gravity data.The numerical experiments show that the new algorithm can efficiently exploit different geophysical data to resolve underground structures.Compared to single geophysical inversion,the problem with the non-uniqueness in geophysical inversions is largely improved using this appraoch.
展开▼
