Using of local indicators of spatial association for evaluation of spatial accuracy of DEM

Using of local indicators of spatial association for evaluation of spatial accuracy of DEM
Jana Svobodova, JakubMirijovsky, Ales Vavra, Jan Brus and Helena Kilianova Palacky 

University in Olomouc, tr. Svobody 26, Olomouc, 771 46 (j.svobodova@upol.cz, jakub.mirijovsky@upol.cz, ales.vavra@upol.cz, jan.brus@upol.cz, helena.kilianova@upol.cz)

Abstract: In the study, whose results are presented in the paper, the statistical method LISA (local indicators of spatial association) has been used for the delimitation of the areas with the statistical significant error values. The error values obtained by subtraction of estimated values of altitude (DEM) from the reference surface were used as the input data for the local spatial cluster analysis. The aim of an interpretation of the results of local cluster analysis was to determine the behaviour of errors (location and size) from the perspective of different quality DEMs, i.e. DEMs created by different interpolation methods and their settings. The main significance of LISA lies in the spatial expression that allows you to monitor the extent, distribution and overall structure of clusters. When comparing the results derived from selected high-quality and low-quality DEM or DEM created by the different interpolation methods it is necessary to realize the different importance of these three aspects. This was reflected in the creation of rules for interpretation and evaluation the results of local cluster analysis, which is one of the main results of the study.

Keywords: DEM, LISA, clusters, statistical significance, spatial accuracy.

AttachmentSize
SvobodovaAccuracy2012.pdf484.7 KB