Geoscience Reference
In-Depth Information
Fig. 14.3 Examples of two DEMs (area of Orlovská plošina) with different properties and
examples of hammock plots created from previous DEMs (Orlovská plošina) 2
n
1
n
2
RMSE Z =
1 (
Z di
Z ri )
(14.1)
i
=
where Z di is the i-th value of altitude from the DEM surface, Z ri is the corresponding
original altitude, and n is the number of checked points.
A higher RMSE value means a higher spread between two data sets; an ideal
value should not go over half of the value of the interval of the original contour
lines. The main advantage of RMSE is the simplicity of the calculation and its clear
concept. However, this index represents a simple global degree of deviations, and
therefore is not able to fully clarify spatial changes of errors over the interpolated
surface. To better understand and quantify DEM inaccuracies, surfaces of the inac-
curacies are used more often, especially the hammock plot (primarily when the input
data are based on contour lines).
The Hammock plot is the surface that originates by whole-number division of
values of DEM, the divisor being the value of the basic interval of the input con-
tour lines. Thus, it is possible to detect errors arising from interpolation from
contour lines (or possibly points generated from contour lines), which means
Search WWH ::




Custom Search