Clustering Analysis of Gravity Data to Compute Local Geoid

In this paper, the clustering analysis method is used in the optimization of observed gravity data during the computation of local geoid, while the two slopes of terrain data can be utilized as criteria. Further, a numerical experiment is carried out in hill area. Compared with the result from non-deleted observed gravity data, the deleted data can still work out acceptable result of similar precision. The numerical experiment validates the feasibility of clustering analysis method which provides a new approach to optimize observed gravity data for computing local geoid.