中文摘要：

针对现有空间插值方法对样点空间分布及结构约束考虑较少,难以保真原有空间数据的统计参量等问题,利用Voronoi和Delaunay的相互关系,建立了基于样点分布V-邻域结构的插值控制点自适应生成方法,构建了顾及样点分布结构与分布密度的结构保持空间插值方法。基于中国气象台站日均气温数据的方法验证与对比表明,相比于常用的空间插值算法,本文算法具有更好的结构自适应性,且对原始数据的空间统计特征具有更好的保持性。

英文摘要：

Existing spatial interpolation methods make little consideration of the irregular spatial distribution and structural constraints of sampling points, and rarely maintain the accuracy of spatial statistical parameters. In this paper, the correlation between Voronoi and Delaunay is used to construct a self-adaptive control point generating method based on the V-neighborhood structure of sampling points. On this basis, a structure-preserving interpolation method, which takes the structure constraints of spatial and density distribution of sample points into consideration, was established. The method was validated with Chinese weather station network data. The results suggest that the algorithm has better structural self-adaptability and maintains more precisely statistical spatial features as compared with the commonly used spatial interpolation algorithms.