卫星重力梯度数据的网格化方法研究

STUDY ON GRIDDING METHODS OF SATELLITE GRAVITY GRADIENT DATA

利用模拟的卫星重力梯度数据,对移动平均法、反距离加权平均法、加权移动趋势面法和普通克里金法的网格化效果进行了分析和比较,结果表明:搜索半径的选择,既要考虑到能够包含足够多的信息来反映原始数据的空间分布特性,又要考虑到计算速度的要求;离散数据的网格化,必须考虑已知点与待求点之间的距离对网格化结果的影响;从网格化结果的精度、计算速度以及对原始数据的逼近效果来看,反距离加权平均法最适用于离散的卫星重力梯度数据的网格化。

Four gridding methods： the moving average method,the inverse distance weighting method,the weighted moving trend surface method and the ordinary Kriging method,which are in common use to deal with the scattered data,are introduced.The gridding effects of the four methods are analyzed and compared with each other by using simulated satellite gravity gradient data.The results show that there should be enough information to be included to reflect the spatial distribution characteristics of the original data and the requirement of computational velocity in the choice of the searching radius;the influence of the distance between the known points and the computational points must be considered in the gridding of the scattered data.It is shown that the inverse distance weighting method is suitable for gridding the scattered satellite gravity gradient data in view of the accuracy of the gridding results,the computational velocity and the approaching effect to the original data.