二维直流电阻率与射频大地电磁双参数聚类联合反演

考虑到射频大地电磁（RMT）法反演介电常数具有可行性，但目前RMT与直流电阻率（DC）联合反演研究中均只进行了电阻率参数的反演，本文基于模糊C均值（FCM）聚类算法，开展了RMT与DC的电阻率与介电常数联合反演.首先，检验了介电常数对RMT观测的影响，并通过DC与RMT法的电阻率参数反演，分析了联合反演相较于单一方法反演的优势；基于双数据集反演电阻率的优势，辅以FCM聚类的双参数联合约束，进一步实现最佳介电常数模型反演.理论试算表明，DC与RMT法的电阻率联合反演，可补足RMT浅部高阻分辨率不理想的缺点，同时改善DC法深部反演能力的不足；此外，通过FCM聚类的联合约束以及真实物性的引导，相较于单一RMT方法，DC与RMT的联合反演可恢复更准确的地下介电常数分布.

Dual-parameter clustering joint inversion of two-dimensional direct current resistivity and radio-magnetotelluric data

The permittivity inversion of radio-magnetotelluric (RMT) is available, but only resistivity parameter is inverted in the joint inversion of RMT and direct current (DC) resistivity in current study. In this paper, a joint inversion of resistivity and permittivity using RMT and DC data is carried out based on fuzzy C-means (FCM) clustering technology. Firstly, the influence of permittivity on RMT observation is tested, and the advantages of joint inversion over the single inversion are analyzed through the resistivity parameter inversion of DC and RMT methods. Based on the advantages of resistivity inversion using dual data sets, the optimal permittivity model is obtained coupled with the dual-parameter mutual constraint of FCM clustering. Synthetic examples show that, the resistivity inversion combined with the DC and RMT methods not only refine the resolution of resistive structure in the shallow subsurface of RMT, but also improve the deep inversion ability of the DC method. In addition, the joint inversion of DC and RMT can recover more accurate permittivity distribution than the single RMT method through the mutual constraint of FCM clustering and the guidance of truth property.