Kriging法在区域重力场插值中的适用性研究

摘要

由于重力观测网点位分布不均匀,密度不够,在精细刻画区域重力场变化时受到很大限制,因此需采用合理的插值算法.文中介绍Kriging插值算法和变异函数理论模型的基础上,结合陕西地区2011—2012年离散重力变化数据,利用最小二乘法得到各个理论变异函数模型的拟合参数,并将实验变异函数模型用于Kriging插值算法,同时综合考虑交叉验证精度统计、插值精度统计结果,以此来研究Kriging插值算法的适用性,研究结果表明:基于最小二乘法获得的球形模型、指数模型拟合参数精度高,高斯模型、对数模型精度最差;基于球形模型、指数模型的实验变异函数用于Kriging插值算法得到的插值结果精度最高、图像平滑自然、异常区明显,是一种适合陕西地区重力离散数据进行插值计算的有效算法.%Because the gravity observation network distribution is not uniform and the point density is not enough,it is very limited to describe the variation of regional gravity field,w hich needs to use the reasonable interpolation algorithm.Based on the detailed introduction of the Kriging interpolation algorithm and the variation function theory,this paper unifies the Shaanxi area 2011 ~ 2012 discrete variation of gravity data,obtains the fitting parameters of theoretical variation function model by least square method,and uses the experimental variation model for Kriging interpolation algorithm.Considering the cross validation accuracy statistics,the interpolation precision results,the application to study Kriging interpolation algorithm,the result show s that:the fitting parameter accuracy of spherical model and exponential model are high based on the least square method,and the accuracy of Gauss model and logarithmic model is the worst;the experimental variation function based on the exponential model and spherical model used in the Kriging interpolation algorithm,can get the highest accuracy of interpolation, meanw hile the image is smooth and the abnormal area is obvious,so it is a suitable interpolation calculation algorithm for gravity discrete data of Shaanxi area.