3.3. Error analysis and quality of extracted DEMs
As pointed out repeatedly in previous works (e.g. Wechsler, 2007
and references therein), errors and resulting uncertainty can never be
fully eliminated in spatial data. Therefore, the quality and accuracy of
all individual DEMs as well as subtracted differential DEMs were
evaluated carefully. Results of error analyses, with a brief introduction
to the methods of evaluation, are outlined below.
The quality of spatial data involves several attributes such as
precision, accuracy, and reliability. Despite frequent usage, this
terminology is not unified. We use the terminology of Karel et al.
(2006) in the following text, namely “precision” measures how well
the model fits the input data and “accuracy” characterizes its
conformity with external data. Both the precision and the accuracy
are evaluated by parameters with no spatial aspects such as the mean
error (ME), standard deviation (SD), or RMSE characterizing the global
quality of DEM (according to Karel and Kraus, 2006)