Özet:

   The accuracy of digital elevation models which are constituted in accordance with different parameters, are quite important in terms of spatial analysis and representation of land topography familiar to reality. The purpose of this study is to research the positional accuracy of different interpolation methods (Inverse Distance Weighted (IDW), Ordinary Kriging (OK), Modified Shepard's (MS), Multiquadric Radial Basis Function (MRBF) and Triangulation with Linear (TWL)) statistically which are used for forming digital elevation models in terms of different point densities (100% and 50%), different point distributions (curvature, grid, uniform and random), and land types with different topography (flat, hilly, mountainous). With this aim, analysis of variance (ANOVA) was implemented. For sub-groups obtained from point densities, point distributions and land types, it was sought statistically with ANOVA whether there was a significant difference amongst the positional accuracies obtained from interpolation methods. The accuracy reaching from decimeters to meters was obtained from especially mountainous and hilly land types. The land topography affects the accuracy of formed digital elevation models. The results obtained from all (100%) data sets were always found better than 50% data sets. Generally speaking, similar properties resulted for curvature and uniform distributions in different point distributions (50%), and better accuracies were obtained considering other distributions. Grid distribution resulted worse than other distributions. OK and MRBF interpolation methods showed similar properties for different land types and point distributions.  The accuracy of these methods resulted better considering other methods. The accuracy obtained from IDW method for all three land types was less than other methods. The accuracy obtained from MS and TWL methods was similar. The interpolation methods used in this study were ordered in terms of obtained accuracy from the best to the worst as MRBF, OK, MS, TWL and IDW.

Anahtar Kelimeler: