Spatial Interpolation Method of Scalar Data Based on Raster Distance Transformation of Map Algebra

Spatial interpolation of scalar data is widely used in some fields such as partition of biological type distribution and land-use planning. Usually, this kind of spatial interpolation is based on the nearest neighborhood interpolation. In this paper, some shortages of traditional spatial interpolation method based on amalgamation of voronoi polygons, such as non intelligence, large consumption computation, and non-expansibility to multi-dimension, are analyzed. In map algebra, raster square plane is considered as metric space for raster distance transformation, which could make distance transformation more precise and easily extend to multi-dimension. Based on raster square plane, raster distance transformation can record not only the spatial distance between any raster and it´s nearest neighborhood pixel but also classification attributes of the nearest neighbor entities of the raster. Through extracting the raster that the distance data and the classification data are symmetrical and similarly symmetrical on its eight directions, the spatial distribution boundary of scalar data can be drawn up and the spatial interpolation of scalar data is achieved. At last, the benefits and applications of this kind of spatial interpolation are analyzed. This algorithm has advantages of the extendibility and easy initialization, which can be widely used for the space subdivision such as geographic divisions of biological types, analysis of facilities location.