Figure 9. The location of the study area on the trapezoids of soviet topographic maps Each reference point is located, as a rule, randomly, and the standard deviation of discretisation error for each pixel of the raster model is calculated by formula 2. Methodology Figure 11. Sheets of soviet topographic maps for the study area Figure 10. DEM of the Kursk region obtained as a result of the SRTM mission Figure 12. Scanned layers of elevation data the classification by HSV values is performed. To do this, we manually select ranges for each of the three HSV pa- rameters, that correspond to the color of the contour lines, by analyzing a specific scan of the topographic map. A map of contour lines is created, in which the values of 1 are assigned to pixels that fall into the ranges for all HSV parameters, and the remaining pixels are assigned the value 0. As a result of the above steps, we automatically get the contours shown in Figure 12, which also shows the rest of elevation information, which was obtained by The ANUDEM algorithm is implemented in ArcGIS software as a Topo to Raster method, which was used for building the DEM. For comparison, DEM was created us- ing other 12 building methods, which are implemented in ArcGIS and Surfer software products: Figure 13. DEMs built using different interpolation methods Table 1. Confidence interval boundaries where X; is the elevation value obtained from the inter- polated DEM, X: is the elevation value obtained from the Figure 14. Check points layout A description of the methods implemented in Surfer can be found, for example, in Yang et al. (2004). Figure 15. The results of estimating the accuracy of DEM interpolation methods
