MM
5
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B4. Istanbul 2004
— generation of the Multivariable Digital Terrain Model
-MDTM- that includes all the variables that has been
extracted from the statistical (and geostatistical)
analysis.
— non-supervised classification (ISODATA method) of
the MDTM.
— analysis of the detected classes characteristics.
Quality control of the classification.
DIGITAL ELEVATION MODEL
— classification of the experimental data considering the
window where each point is included.
— basic statistics and directional variograms of the
classified data.
Moving windows schema parameters
(size, spacing, overlap, ...)
E DEM of the windows
Statistics (mean, std.dev.,minimum,quartiles maximum,range....)
Directional variograms
Y
MULTIVARIABLE DIGITAL
TERRAIN MODEL
W 32 variables
ss Statistical variables (mean, variance, CV, …)
Variograms parameters (anisotropy ratio, variogram values...)
ISODATA Unsupervised Classification
v
CLASSIFIED MAP
v
CLASSIFIED DATA
— se Ciassification Statistics
m Variograms
Class Statistics
Figure 1. Workflow of the proposed methodology
3. EXAMPLE
3.1 Data set
In order to show the capabilities of the proposed
methodology, it has been applied to an example area. The
selected area is located between the Granada and Almería
province (S of Spain) being banked to the S for Sierra
Nevada's relief (figure 2).
The used information for this terrain morphological
classification has been a digital elevation model in format
grid with 20x20m's spacing generated starting from a contour
map's digitalization to scale 1:50000.
The selected area was selected for its geological interest. In
the area, there is a important diversity of morphological
units. In the south, the high summit of Sierra Nevada's
foothills with a very important drainage network
development is present. In the north the arca the terrain
presents minor heights (Negratin Dam zone) A very
important surface (central area) is occupied by the Guadix
Basin that shows up as a flat area surrounded of important
relief.
The zone has 50x70 km surface that produces a DEM with
8,756,001 points distributed according to a mesh 2501
samples x 3501 lines. In figure 3 a DEM grayscale
representation is shown. In table | the most important basic
statistics of the terrain heights are summarized.
Figure 2. Geographical localization
The most important characteristic of the area is the amplitude
of the heights range that stretches out from 500 meters on the
sea level to the 3480 meters, (Mulhacen peak, Iberian
Peninsula maximum height). The mean value is 1261 meters
and the median value 1101 meters (that is representative of
the heights of the Guadix plain that occupies an important
surface of the selected area).