Solution for the Robust Estimation of Heterogeneous Data Fusion Based on Classification Estimation

Abstract

Data fusion is an important issue as multi-source observation technology is widely used in geoscience. Although the problem of robust estimation exists widely in the data fusion, few solutions have been reported. This paper investigates a solution for the robust estimation of heterogeneous data fusion implementing classification, robust estimation, and data fusion. A new approach based on Msplit estimation is constructed to define accurate scales for multi-source observation data and, adopting the Institute of Geodesy and Geophysics (IGG)III weight function, an iterative algorithm is proposed for the above problem. Finally, two instances are considered in verifying the feasibility of the presented solution.