The Comparison And Optimization Of Spatial Interpolation Approaches

In the atmosphere, elevation, mining and other geographic sample survey areas, onlydiscrete data can be achieved due to limitation of means of observation, the data beyondthe scope of the observation sites have to be attained through spatial interpolation methodbased on the existed data. In order to get more detailed, accurate and intuitive interpolationmap results, and assist the scientific decision-making of management and technical staff,improve efficiency, reduce costs, improve the interpolation accuracy is the main focus ofthe study of spatial interpolation. Nevertheless, most current papers are focusing on theapplication of spatial interpolation other than their optimization, or on the optimization ofinterpolation models. In fact, it is necessary to optimize the selecting points methods ofspatial interpolation which affect the accuracy of interpolation a lot. In this paper, the studyis based on the monthly average precipitation data of Fujian Province in December2005.Through the comparison of conventional selecting points methods and Voronoidiagram-based selecting points method, the latter is proved to be with more interpolationaccuracy. The main content is listed as below:1. There is an overview summary of application areas, research history and the latestadvancements of spatial interpolation methods, along with a classification ofinterpolation methods based on spatial data organization，and detailed description ofprinciples, advantages&disadvantages, exclusions，the research status quo of variousinterpolation methods.2. Overall interpolation, spline function method, inverse distance weighting method,Kriging interpolation were used on the experimental zone data, the results show thatthe overall interpolation method is suitable for general trend analysis, spline functionmethod can show local details, the inverse distance weighting method is prone toproduce "bull eye" phenomenon, Kriging is suitable for data with a higher degree ofspatial autocorrelation. Further possible approach of optimization of spatialinterpolation can be obtained from the perspectives of interpolation model andselecting points methods.3. Voronoi graph is generated by experimental zone data through the secondarydevelopment on ArcGIS desktop series, which is used for selecting points in the interpolation process. Inverse distance weighting method is used on the interpolationof experimental area; comparing with the interpolation results of Voronoi Graph basedselecting points method and conventional selecting points method, accuracy analysisshows that the research results of the paper have practical value in the optimization ofspatial interpolation.