摘要：【中文】本发明公开了一种基于信任传播的气象地面要素插值方法，具体涉及气象数据插值领域，具体方法步骤如下：S1、设置概率密度核函数，计算站点密度；S2、增加虚拟站点；S3、使用k‑NN Barnes二分因子图算法建立因子图；S4、kNN梯度检验；S5、引入数值高程数据，使用信任传播算法，在站点‑站点因子图的基础上对站点V进行气象要素插值计算；S6、在站点‑格点场因子图的基础上，利用站点V气象要素值，对格点场中各格点进行气象要素插值计算；S7、计算风速变化量。本发明通过建立站点‑站点因子图、站点‑格点场因子图，实现快速多要素的离散站点到格点场的插值；通过包含DEM数据的信任传播算法进行插值，提高各气象要素插值结果的准确率。 【EN】The invention discloses a meteorological ground element interpolation method based on belief propagation, and particularly relates to the field of meteorological data interpolation, wherein the method comprises the following specific steps: s1, setting a probability density kernel function, and calculating the site density; s2, adding virtual sites; s3, establishing a factor graph by using a k-NN Barnes bipartite factor graph algorithm; s4, performing kNN gradient test; s5, introducing numerical elevation data, and performing meteorological element interpolation calculation on a site V on the basis of a site-site factor graph by using a belief propagation algorithm; s6, on the basis of the site-grid point field factor graph, performing meteorological element interpolation calculation on each grid point in the grid point field by using the site V meteorological element value; and S7, calculating the wind speed variation. The invention realizes fast multi-element discrete site-grid point field interpolation by establishing a site-site factor graph and a site-grid point field factor graph; interpolation is carried out through a trust propagation algorithm containing DEM data, and the accuracy of interpolation results of all meteorological elements is improved.