一种基于数据融合的地球物理数据联合反演方法——以VES和MT为例

A GEOPHYSICAL INTEGRATION INVERSION METHOD BASED ON DATA FUSION

从地球物理反演的基本概念出发 ,认为地球物理反演是对实测数据的地球物理属性的理解或解释 ,多方法的地球物理数据联合反演是一种多传感器的数据融合 .本文分析了地球物理数据的模糊特性 ,采用基于语义的模糊化方法 ,使不同物理意义和尺度的特征数据及测区的地质和地球物理背景成为一体 ,结合地球物理专家解释的方法 ,利用基于模糊逻辑系统的神经网络实现了融合 .该方法充分利用了各种地球物理探测数据的全部信息 ,避免了线性反演的复杂计算 ;其数据融合的观点 ,为解决地球物理联合反演问题提供了新的思路 .通过模拟实验和应用实例验证 。

The geophysical inversion can be regarded as an understanding or interpretation of geophysical property on the basis of the observed data, while the inversion of multiple geophysical data can be regarded as a fusion of multiple sensor data. The fuzzy features of geophysical data were analyzed. The characteristic data of different physical meaning and scales, and the knowledge of geology and geophysics in the studied area are combined using the method of a uniform linguistic defuzzification. The method integrates the geophysical expert knowledge, and carried out data fusion using fuzzy inferential network. The method makes full use of all information of manifold geophysical data, avoiding the complicated calculation in linear joint inversion, and provides a new idea for the geophysical inversion about the data fusion point of view. It was proved to be effective through modeling test and application in Guyuan.