银川市红墩子矿区钻孔亚粘土数据kriging插值方法比较研究  

Comparison of kriging interpolation methods based on the drilling data of hongdunzi mining area in Yinchuan city

作　　者：孙涵 SUN Han(School of Information Engineering,China University of Geosciences,Beijing 100083,China)

机构地区：[1]中国地质大学信息工程学院,北京100083

出　　处：《河北省科学院学报》2020年第3期55-64,共10页Journal of The Hebei Academy of Sciences

基　　金：自然资源部信息中心专项课题资助(3-4-2019-174)。

摘　　要：利用银川市红墩子矿区钻孔亚粘土数据,在ArcGIS支持下,使用四种克里金插值方法对亚粘土分布高程进行预测。采用交叉验证对插值精度进行评价时,根据最优条件对五参数进行线性相加作为判断预测误差大小的辅助工具,并综合插值结果的统计特征和分布图,得到如下结论:(1)针对本次研究数据,使用协同克里金插值方法,半变异函数模型为高斯函数,步长为8插值效果最好。(2)无论哪种插值方法,使用高斯半变异函数模型都是最优的。(3)交叉验证时将五个参数根据最优条件进行线性相加,能够在一定程度上反映预测误差好坏。(4)在研究数据较少,但它与其他数据存在较大相关性时可以使用协同克里金法,会在较大程度上提高插值精度。With the aid of ArcGIS,four kinds of kriging interpolation methods are used to predict the distribution elevation of mild clay based on the drilling data of hongdunzi mining area in Yinchuan city.When the interpolation accuracy is evaluated by cross validation,the five parameters are added linearly according to the optimal conditions as the auxiliary tool to judge the prediction error.The following conclusions are obtained by synthesizing the statistical characteristics and distribution diagram of the interpolation results:(1)According to the data of this study,using the co-kriging interpolation method,the semi variogram model is Gaussian function,and the step size of 8 interpolation is the best.(2)No matter which interpolation method,the Gaussian semi variogram model is the best.(3)In cross validation,five parameters are added linearly according to the optimal conditions,which can reflect the prediction error to a certain extent.(4)Co-kriging can be used when there is less research data,but it has a large correlation with other data,which will greatly improve the interpolation accuracy.

关 键 词：克里金插值 岩石高程分布 红墩子矿区 交叉验证判断公式 ARCGIS 

分 类 号：P634[天文地球—地质矿产勘探]