论文部分内容阅读
时空Kriging算法的核心,是将变差函数的概念扩展到时空域,变差函数的构建过程基础是计算时间切片和空间切片的向量距离。该文讨论了向量距离对构建时空变差函数的影响,提出了空间距离加角度差异的向量距离模型。以地磁场观测数据作为对象,分别用L1范数、L2范数和新距离模型对数据进行分析,比较3种距离定义下的时空Kriging插值性能。结果表明:加入了角度信息的向量距离,能够更有效地表征数据,提高时空Kriging的插值精度。
The core of spatio-temporal Kriging algorithm is to extend the concept of variogram to spatio-temporal domain. The construction of variogram is based on calculating the vector distance between time slice and space slice. This paper discusses the influence of vector distance on constructing spatio-temporal variation function and proposes a vector distance model of spatial distance plus angle difference. Taking geomagnetic observational data as object, the data were analyzed by L1 norm, L2 norm and new distance model respectively. The space-time Kriging interpolation performance under three kinds of distance was compared. The results show that adding the vector distance of angle information can characterize the data more effectively and improve the interpolation precision of space-time Kriging.